Introduction to Boosting

This module introduces the concept of boosting, an ensemble learning technique in which weak learners are combined together by using weights. This iterative process is continued by modifying the weights to emphasize inaccurate predictions, which can then be better predicted in future iterations.

Objectives

By the end of this lesson, you will be able to

explain the concept of boosting,
articulate the benefits of using boosting, and how it differs from bagging, and
apply boosting algorithms, such as the gradient boosted decision tree

Boosting

The decision tree is one of the simplest machine learning algorithms to understand. Simple to develop and easy to apply, they quickly became a popular machine learning tool. However, they can be prone to overfitting the data, especially when they become too large. As a result, ensemble techniques were developed that combined the predictions from many decision trees, that in general are small and might be trained only on a fraction of the data, to make a more powerful meta-estimator.

In the Bagging lesson, we explored the concept of bagging, which can be used to construct one type of ensemble estimator. Bagging employs bootstrap aggregation to first create samples of the original data set, which are each used to construct a weak learner, generally by using a decision tree. Finally, these weak learners are aggregated together to make a more powerful predictor.

In this module, we introduce a second approach, known as boosting, where many weak learners, for example, very shallow decision trees (also known as tree stumps), are trained. The predictions from these weak learners are combined together, for example, by majority voting, and the results are used to generate weights for the input training data. These weights are used by the boosting algorithm to emphasize instances that are incorrectly predicted, and to deemphasize instances that are correctly predicted. This process continues iteratively, during which the weak learners eventually become more accurate.

I introduce two algorithms that implement boosting: Gradient Tree Boosting and Adaboost. For space considerations, we focus most of the module on effectively using Gradient Boosted Trees (GBT). First, we demonstrate this algorithm on the Iris classification task, which also allows us to demonstrate how this algorithm can be used to compute feature importances, as well as use decision surfaces to explore different hyperparameters. Finally, the Adaboost algorithm is quickly presented by applying it to these same data sets.

Throughout this module, you should think about how these results from boosting are different than the results from the other algorithms, in particular the bagging algorithms introduced in a different module. You also should explore how changes in the default hyperparameter values affect the performance of this algorithm.

Formalism

While there are a number of different boosting algorithms two of the more popular ones are the Adaboost algorithm, which is short for adaptive boosting, and the Gradient Tree Boosting algorithm. The Adaboost algorithm works as the earlier boosting description outlined. Weights are generated in each iteration and used to improve the overall prediction of the aggregated weak learners. The Gradient Tree Boosting algorithm extends this to support arbitrary cost (or loss) functions, beyond the minimization of the deviation between predicted and actual labels (classification), or least squares (regression).

These two algorithms will both be explored in this module, although most of the module focuses on the Gradient Tree Boosting, with the Adaboost algorithm presented at the end for comparison purposes. The construction of the individual base learners is based on a permutation of the features when determining the best split. Thus, the individual learners can be different, even if they are constructed by using all of the training data. These and other characteristics of the Gradient Tree Boosting algorithm are controlled by a set of hyperparameters, many of which are the same as for bagging estimators. Some of the more useful hyperparameters include:

n.trees

Integer specifying the total number of trees to fit. This is equivalent to the number of iterations and the number of basis functions in the additive expansion. Default is 100.

interaction.depth

Integer specifying the maximum depth of each tree (i.e., the highest level of variable interactions allowed). A value of 1 implies an additive model, a value of 2 implies a model with up to 2-way interactions, etc. Default is 1.

n.minobsinnode

Integer specifying the minimum number of observations in the terminal nodes of the trees. Note that this is the actual number of observations, not the total weight.

shrinkage

a shrinkage parameter applied to each tree in the expansion. Also known as the learning rate or step-size reduction; 0.001 to 0.1 usually work, but a smaller learning rate typically requires more trees. Default is 0.1.

Classification: Iris Data

Here we go again 😆

library(tidyverse)

set.seed(1)
#lets split the data 60/40
library(caret)
trainIndex <- createDataPartition(iris$Species, p = .6, list = FALSE, times = 1)

#grab the data
irisTrain <- iris[ trainIndex,]
irisTest  <- iris[-trainIndex,]

ggplot(data=irisTrain)+geom_point(mapping = aes(x=Petal.Length,y=Petal.Width,color=Species),alpha=0.5) + labs(color = "Training Species")+
geom_point(data=irisTest, ,mapping = aes(x=Petal.Length,y=Petal.Width,shape=Species)) + labs(shape = "Testing Species") +
  ggtitle("The data")+
  theme(plot.title = element_text(hjust=0.5, size=10, face='bold'))

The Model

set.seed(1)
#added something here
IrisXGB<- train(
  form = factor(Species) ~ .,
  data = irisTrain,
  #here we add classProbs because we want probs
  trControl = trainControl(method = "cv", number = 10,
                           classProbs =  TRUE),
  method = "gbm",
  tuneLength = 20,
  #add this please
  verbose=FALSE)

#IrisXGB
knitr::kable(IrisXGB$bestTune)

	n.trees	interaction.depth	shrinkage	n.minobsinnode
142	100	8	0.1	10

plot(IrisXGB)

IrisXGB_Pred<-predict(IrisXGB,irisTest,type="prob")

knitr::kable(IrisXGB_Pred)%>%
  kableExtra::kable_styling("striped")%>%
  kableExtra::scroll_box(width = "50%",height="300px")

setosa	versicolor	virginica
0.9999999	0.0000001	0.0000001
1.0000000	0.0000000	0.0000000
1.0000000	0.0000000	0.0000000
0.9999999	0.0000000	0.0000000
0.9999999	0.0000001	0.0000001
0.9999604	0.0000391	0.0000005
0.9999998	0.0000002	0.0000000
0.9998762	0.0001231	0.0000007
0.9999996	0.0000004	0.0000000
1.0000000	0.0000000	0.0000000
0.9999999	0.0000000	0.0000001
0.9999998	0.0000002	0.0000000
0.9999999	0.0000000	0.0000000
0.9999993	0.0000007	0.0000000
1.0000000	0.0000000	0.0000000
0.9956348	0.0023072	0.0020580
0.9999999	0.0000001	0.0000001
0.9999996	0.0000004	0.0000000
0.9999995	0.0000004	0.0000001
1.0000000	0.0000000	0.0000000
0.0000146	0.2480013	0.7519840
0.0000000	0.9999999	0.0000000
0.0000000	1.0000000	0.0000000
0.0000004	0.9999952	0.0000044
0.0000001	0.9999883	0.0000116
0.0000000	0.9999998	0.0000001
0.0000000	1.0000000	0.0000000
0.0000001	0.9999947	0.0000053
0.0000003	0.0000695	0.9999302
0.0000000	1.0000000	0.0000000
0.0000007	0.9994671	0.0005322
0.0000001	0.0000253	0.9999747
0.0000001	0.9999972	0.0000028
0.0000001	0.9999976	0.0000023
0.0000001	0.0000211	0.9999788
0.0000709	0.9999073	0.0000218
0.0000101	0.9966311	0.0033587
0.0000000	0.9999997	0.0000003
0.0000001	0.9999953	0.0000046
0.0000001	0.9999996	0.0000004
0.0000000	0.0000017	0.9999983
0.0000000	0.0000106	0.9999893
0.0000000	0.0000017	0.9999983
0.0000000	0.0000017	0.9999983
0.0000000	0.0000038	0.9999962
0.0000002	0.0002156	0.9997843
0.0000000	0.0000008	0.9999992
0.0000000	0.0000001	0.9999999
0.0000015	0.0245239	0.9754746
0.0000009	0.0048799	0.9951192
0.0000004	0.0001712	0.9998285
0.0000099	0.1688174	0.8311727
0.0000009	0.0007682	0.9992310
0.0000000	0.0000017	0.9999983
0.0000000	0.0000034	0.9999966
0.0000000	0.0000031	0.9999969
0.0000000	0.0000040	0.9999960
0.0000000	0.0000013	0.9999987
0.0000000	0.0000013	0.9999987
0.0000002	0.0003979	0.9996019

irisXGBtestpred<-cbind(IrisXGB_Pred,irisTest)

irisXGBtestpred<-irisXGBtestpred%>%
  mutate(prediction=if_else(setosa>versicolor & setosa>virginica,"setosa",
                            if_else(versicolor>setosa & versicolor>virginica, "versicolor",
                                    if_else(virginica>setosa & virginica>versicolor,"virginica", "PROBLEM"))))

table(irisXGBtestpred$prediction)


    setosa versicolor  virginica 
        20         16         24

XgbConfusion<-confusionMatrix(factor(irisXGBtestpred$prediction),factor(irisXGBtestpred$Species))

XgbConfusion

Confusion Matrix and Statistics

            Reference
Prediction   setosa versicolor virginica
  setosa         20          0         0
  versicolor      0         16         0
  virginica       0          4        20

Overall Statistics
                                         
               Accuracy : 0.9333         
                 95% CI : (0.838, 0.9815)
    No Information Rate : 0.3333         
    P-Value [Acc > NIR] : < 2.2e-16      
                                         
                  Kappa : 0.9            
                                         
 Mcnemar's Test P-Value : NA             

Statistics by Class:

                     Class: setosa Class: versicolor Class: virginica
Sensitivity                 1.0000            0.8000           1.0000
Specificity                 1.0000            1.0000           0.9000
Pos Pred Value              1.0000            1.0000           0.8333
Neg Pred Value              1.0000            0.9091           1.0000
Prevalence                  0.3333            0.3333           0.3333
Detection Rate              0.3333            0.2667           0.3333
Detection Prevalence        0.3333            0.2667           0.4000
Balanced Accuracy           1.0000            0.9000           0.9500

Lets have a little fun with ggplot and the confusion matrix 😈

ggplot(as.data.frame(XgbConfusion$table))+ 
  geom_raster(aes(x=Reference, y=Prediction, fill=Freq)) + 
  geom_text(aes(x=Reference, y=Prediction, label=Freq)) +
   scale_fill_gradient2( low = "darkred", high = "pink", na.value="black", name = "Freq" )+
  scale_x_discrete(name="Actual Class") + 
  scale_y_discrete(name="Predicted Class")+
  ggtitle("Confusion is fun")+
  theme(plot.title = element_text(hjust=0.5, size=10, face='bold'))

How did I figure out how to make the above graph? 🤔

Gradient Tree Boosting: Feature Importance

We can leverage the fact that the gradient tree boosting algorithm builds many weak learners and iteratively improves them to determine the importance of the different features. By default, these values simply encode the importance of a feature, where higher values are more important. We can take the ratio of these values, however, to compute a relative importance.

library(gbm)

summary(IrisXGB)

                      var   rel.inf
Petal.Length Petal.Length 54.367434
Petal.Width   Petal.Width 38.353661
Sepal.Width   Sepal.Width  5.463525
Sepal.Length Sepal.Length  1.815380

#had to add something 
V<-caret::varImp(IrisXGB, n.trees=500)$importance%>%
  arrange(desc(Overall))

knitr::kable(V)

	Overall
Petal.Length	100.000000
Petal.Width	69.527789
Sepal.Width	6.941965
Sepal.Length	0.000000

ggplot2::ggplot(V, aes(x=reorder(rownames(V),Overall), y=Overall)) +
geom_point( color="blue", size=4, alpha=0.6)+
geom_segment( aes(x=rownames(V), xend=rownames(V), y=0, yend=Overall), 
color='skyblue') +
xlab('Variable')+
ylab('Overall Importance')+
theme_light() +
coord_flip()

Gradient Tree Boosting: Decision Surface

set.seed(1)
#lets split the data 60/40
library(caret)
trainIndex <- createDataPartition(iris$Species, p = .6, list = FALSE, times = 1)

#grab the data
train <- iris[ trainIndex,]
test  <- iris[-trainIndex,]

mygrid<-expand.grid(interaction.depth = seq(2, 11, by = 2),
                        n.trees = seq(50, 250, by = 50),
                        n.minobsinnode = 10,
                        shrinkage = c(0.01, 0.1))

Irisgbm<- train(
  form = factor(Species) ~ .,
  data = train,
  #here we add classProbs because we want probs
  trControl = trainControl(method = "cv", number = 10,
                           classProbs =  TRUE),
  method = "gbm",
  tuneGrid = mygrid,
  verbose=FALSE)

knitr::kable(Irisgbm$bestTune)

	n.trees	interaction.depth	shrinkage	n.minobsinnode
26	50	2	0.1	10

pl = seq(min(iris$Petal.Length), max(iris$Petal.Length), by=0.1)
pw = seq(min(iris$Petal.Width), max(iris$Petal.Width), by=0.1)

# generates the boundaries for your graph
lgrid <- expand.grid(Petal.Length=pl, 
                     Petal.Width=pw,
                     Sepal.Length = 5.4,
                     Sepal.Width=3.1)

IrisgbmGrid2 <- predict(Irisgbm, newdata=lgrid)
IrisgbmGrid <- as.numeric(IrisgbmGrid2)

# get the points from the test data...
testPred <- predict(Irisgbm, newdata=test)
testPred <- as.numeric(testPred)
# this gets the points for the testPred...
test$Pred <- testPred

probs <- matrix(IrisgbmGrid, length(pl), length(pw))

ggplot(data=lgrid) + stat_contour(aes(x=Petal.Length, y=Petal.Width, z=IrisgbmGrid),bins=10) +
  geom_point(aes(x=Petal.Length, y=Petal.Width, colour=IrisgbmGrid2),alpha=.2) +
  geom_point(data=test, aes(x=Petal.Length, y=Petal.Width, shape=Species), size=2) + 
  labs(shape = "Testing Species") +
  geom_point(data=train, aes(x=Petal.Length, y=Petal.Width, color=Species), size=2, alpha=0.75)+
  theme_bw()+ 
  labs(color = "Training Species")+
  ggtitle("Decision Surface")

Exercise 1

Use the tips data set and gradient boosting to predict sex. Include a confusion matrix, roc curve, gain chart, and a lift chart. Look back at the random forest module for the roc, gain, and lift code. 👍 😄

library(curl)

load(curl("https://raw.githubusercontent.com/Professor-Hunt/ACC8143/main/data/tips.rda"))

Gradient Tree Boosting: Regression

We can also use gradient boosting to perform regression; however, in this case we attempt to create trees whose leaf nodes contain data that are nearby in the overall feature space. To predict a continuous value from a tree, we either have leaf nodes with only one feature, and use the relevant feature from that instance as our predictor, or we compute summary statistics from the instances in the appropriate leaf node, such as the mean or mode. The boosting algorithm combines these predictions together to (hopefully) make a more accurate estimator.

Lets get some more interesting data:

# install.packages("devtools")
# devtools::install_github("tyluRp/ucimlr")


knitr::kable(ucimlr::ucidata())%>%
  kableExtra::kable_styling("striped")%>%
  kableExtra::scroll_box(width = "100%",height="300px")

name	type	task	variable_types	observations	variables	year
Abalone	Multivariate	Classification	Categorical, Integer, Real	4177	8	1995
Adult	Multivariate	Classification	Categorical, Integer	48842	14	1996
Annealing	Multivariate	Classification	Categorical, Integer, Real	798	38	NA
Anonymous Microsoft Web Data	NA	Recommender-Systems	Categorical	37711	294	1998
Arrhythmia	Multivariate	Classification	Categorical, Integer, Real	452	279	1998
Artificial Characters	Multivariate	Classification	Categorical, Integer, Real	6000	7	1992
Audiology (Original)	Multivariate	Classification	Categorical	226	NA	1987
Audiology (Standardized)	Multivariate	Classification	Categorical	226	69	1992
Auto MPG	Multivariate	Regression	Categorical, Real	398	8	1993
Automobile	Multivariate	Regression	Categorical, Integer, Real	205	26	1987
Badges	Univariate, Text	Classification	NA	294	1	1994
Balance Scale	Multivariate	Classification	Categorical	625	4	1994
Balloons	Multivariate	Classification	Categorical	16	4	NA
Breast Cancer	Multivariate	Classification	Categorical	286	9	1988
Breast Cancer Wisconsin (Original)	Multivariate	Classification	Integer	699	10	1992
Breast Cancer Wisconsin (Prognostic)	Multivariate	Classification, Regression	Real	198	34	1995
Breast Cancer Wisconsin (Diagnostic)	Multivariate	Classification	Real	569	32	1995
Pittsburgh Bridges	Multivariate	Classification	Categorical, Integer	108	13	1990
Car Evaluation	Multivariate	Classification	Categorical	1728	6	1997
Census Income	Multivariate	Classification	Categorical, Integer	48842	14	1996
Chess (King-Rook vs. King-Knight)	Multivariate, Data-Generator	Classification	Categorical, Integer	NA	22	1988
Chess (King-Rook vs. King-Pawn)	Multivariate	Classification	Categorical	3196	36	1989
Chess (King-Rook vs. King)	Multivariate	Classification	Categorical, Integer	28056	6	1994
Chess (Domain Theories)	Domain-Theory	NA	NA	NA	NA	NA
Bach Chorales	Univariate, Time-Series	NA	Categorical, Integer	100	6	NA
Connect-4	Multivariate, Spatial	Classification	Categorical	67557	42	1995
Credit Approval	Multivariate	Classification	Categorical, Integer, Real	690	15	NA
Japanese Credit Screening	Multivariate, Domain-Theory	Classification	Categorical, Real, Integer	125	NA	1992
Computer Hardware	Multivariate	Regression	Integer	209	9	1987
Contraceptive Method Choice	Multivariate	Classification	Categorical, Integer	1473	9	1997
Covertype	Multivariate	Classification	Categorical, Integer	581012	54	1998
Cylinder Bands	Multivariate	Classification	Categorical, Integer, Real	512	39	1995
Dermatology	Multivariate	Classification	Categorical, Integer	366	33	1998
Diabetes	Multivariate, Time-Series	NA	Categorical, Integer	NA	20	NA
DGP2 - The Second Data Generation Program	Data-Generator	NA	Real	NA	NA	NA
Document Understanding	NA	NA	NA	NA	NA	1994
EBL Domain Theories	NA	NA	NA	NA	NA	NA
Echocardiogram	Multivariate	Classification	Categorical, Integer, Real	132	12	1989
Ecoli	Multivariate	Classification	Real	336	8	1996
Flags	Multivariate	Classification	Categorical, Integer	194	30	1990
Function Finding	NA	Function-Learning	Real	352	NA	1990
Glass Identification	Multivariate	Classification	Real	214	10	1987
Haberman’s Survival	Multivariate	Classification	Integer	306	3	1999
Hayes-Roth	Multivariate	Classification	Categorical	160	5	1989
Heart Disease	Multivariate	Classification	Categorical, Integer, Real	303	75	1988
Hepatitis	Multivariate	Classification	Categorical, Integer, Real	155	19	1988
Horse Colic	Multivariate	Classification	Categorical, Integer, Real	368	27	1989
ICU	Multivariate, Time-Series	NA	Real	NA	NA	NA
Image Segmentation	Multivariate	Classification	Real	2310	19	1990
Internet Advertisements	Multivariate	Classification	Categorical, Integer, Real	3279	1558	1998
Ionosphere	Multivariate	Classification	Integer, Real	351	34	1989
Iris	Multivariate	Classification	Real	150	4	1988
ISOLET	Multivariate	Classification	Real	7797	617	1994
Kinship	Relational	Relational-Learning	Categorical	104	12	1990
Labor Relations	Multivariate	NA	Categorical, Integer, Real	57	16	1988
LED Display Domain	Multivariate, Data-Generator	Classification	Categorical	NA	7	1988
Lenses	Multivariate	Classification	Categorical	24	4	1990
Letter Recognition	Multivariate	Classification	Integer	20000	16	1991
Liver Disorders	Multivariate	NA	Categorical, Integer, Real	345	7	1990
Logic Theorist	Domain-Theory	NA	NA	NA	NA	NA
Lung Cancer	Multivariate	Classification	Integer	32	56	1992
Lymphography	Multivariate	Classification	Categorical	148	18	1988
Mechanical Analysis	Multivariate	Classification	Categorical, Integer, Real	209	8	1990
Meta-data	Multivariate	Classification	Categorical, Integer, Real	528	22	1996
Mobile Robots	Domain-Theory	NA	Categorical, Integer, Real	NA	NA	1995
Molecular Biology (Promoter Gene Sequences)	Sequential, Domain-Theory	Classification	Categorical	106	58	1990
Molecular Biology (Protein Secondary Structure)	Sequential	Classification	Categorical	128	NA	NA
Molecular Biology (Splice-junction Gene Sequences)	Sequential, Domain-Theory	Classification	Categorical	3190	61	1992
MONK’s Problems	Multivariate	Classification	Categorical	432	7	1992
Moral Reasoner	Domain-Theory	NA	NA	202	NA	1994
Multiple Features	Multivariate	Classification	Integer, Real	2000	649	NA
Mushroom	Multivariate	Classification	Categorical	8124	22	1987
Musk (Version 1)	Multivariate	Classification	Integer	476	168	1994
Musk (Version 2)	Multivariate	Classification	Integer	6598	168	1994
Nursery	Multivariate	Classification	Categorical	12960	8	1997
Othello Domain Theory	Domain-Theory	NA	NA	NA	NA	1991
Page Blocks Classification	Multivariate	Classification	Integer, Real	5473	10	1995
Optical Recognition of Handwritten Digits	Multivariate	Classification	Integer	5620	64	1998
Pen-Based Recognition of Handwritten Digits	Multivariate	Classification	Integer	10992	16	1998
Post-Operative Patient	Multivariate	Classification	Categorical, Integer	90	8	1993
Primary Tumor	Multivariate	Classification	Categorical	339	17	1988
Prodigy	Domain-Theory	NA	NA	NA	NA	NA
Qualitative Structure Activity Relationships	Domain-Theory	NA	NA	NA	NA	NA
Quadruped Mammals	Multivariate, Data-Generator	Classification	Real	NA	72	1992
Servo	Multivariate	Regression	Categorical, Integer	167	4	1993
Shuttle Landing Control	Multivariate	Classification	Categorical	15	6	1988
Solar Flare	Multivariate	Regression	Categorical	1389	10	1989
Soybean (Large)	Multivariate	Classification	Categorical	307	35	1988
Soybean (Small)	Multivariate	Classification	Categorical	47	35	1987
Challenger USA Space Shuttle O-Ring	Multivariate	Regression	Integer	23	4	1993
Low Resolution Spectrometer	Multivariate	Classification	Integer, Real	531	102	1988
Spambase	Multivariate	Classification	Integer, Real	4601	57	1999
SPECT Heart	Multivariate	Classification	Categorical	267	22	2001
SPECTF Heart	Multivariate	Classification	Integer	267	44	2001
Sponge	Multivariate	Clustering	Categorical, Integer	76	45	NA
Statlog Project	NA	NA	NA	NA	NA	1992
Student Loan Relational	Domain-Theory	NA	NA	1000	NA	1993
Teaching Assistant Evaluation	Multivariate	Classification	Categorical, Integer	151	5	1997
Tic-Tac-Toe Endgame	Multivariate	Classification	Categorical	958	9	1991
Thyroid Disease	Multivariate, Domain-Theory	Classification	Categorical, Real	7200	21	1987
Trains	Multivariate	Classification	Categorical	10	32	1994
University	Multivariate	Classification	Categorical, Integer	285	17	1988
Congressional Voting Records	Multivariate	Classification	Categorical	435	16	1987
Water Treatment Plant	Multivariate	Clustering	Integer, Real	527	38	1993
Waveform Database Generator (Version 1)	Multivariate, Data-Generator	Classification	Real	5000	21	1988
Waveform Database Generator (Version 2)	Multivariate, Data-Generator	Classification	Real	5000	40	1988
Wine	Multivariate	Classification	Integer, Real	178	13	1991
Yeast	Multivariate	Classification	Real	1484	8	1996
Zoo	Multivariate	Classification	Categorical, Integer	101	17	1990
Undocumented	NA	NA	NA	NA	NA	NA
Twenty Newsgroups	Text	NA	NA	20000	NA	1999
Australian Sign Language signs	Multivariate, Time-Series	Classification	Categorical, Real	6650	15	1999
Australian Sign Language signs (High Quality)	Multivariate, Time-Series	Classification	Real	2565	22	2002
US Census Data (1990)	Multivariate	Clustering	Categorical	2458285	68	NA
Census-Income (KDD)	Multivariate	Classification	Categorical, Integer	299285	40	2000
Coil 1999 Competition Data	Multivariate	NA	Categorical, Real	340	17	1999
Corel Image Features	Multivariate	NA	Real	68040	89	1999
E. Coli Genes	Relational	NA	NA	NA	NA	2001
EEG Database	Multivariate, Time-Series	NA	Categorical, Integer, Real	122	4	1999
El Nino	Spatio-temporal	NA	Integer, Real	178080	12	1999
Entree Chicago Recommendation Data	Transactional, Sequential	Recommender-Systems	Categorical	50672	NA	2000
CMU Face Images	Image	Classification	Integer	640	NA	1999
Insurance Company Benchmark (COIL 2000)	Multivariate	Regression, Description	Categorical, Integer	9000	86	2000
Internet Usage Data	Multivariate	NA	Categorical, Integer	10104	72	1999
IPUMS Census Database	Multivariate	NA	Categorical, Integer	256932	61	1999
Japanese Vowels	Multivariate, Time-Series	Classification	Real	640	12	NA
KDD Cup 1998 Data	Multivariate	Regression	Categorical, Integer	191779	481	1998
KDD Cup 1999 Data	Multivariate	Classification	Categorical, Integer	4000000	42	1999
M. Tuberculosis Genes	Relational	NA	NA	NA	NA	2001
Movie	Multivariate, Relational	NA	NA	10000	NA	1999
MSNBC.com Anonymous Web Data	Sequential	NA	Categorical	989818	NA	NA
NSF Research Award Abstracts 1990-2003	Text	NA	NA	129000	NA	2003
Pioneer-1 Mobile Robot Data	Multivariate, Time-Series	NA	Categorical, Real	NA	NA	1999
Pseudo Periodic Synthetic Time Series	Univariate, Time-Series	NA	NA	100000	NA	1999
Reuters-21578 Text Categorization Collection	Text	Classification	Categorical	21578	5	1997
Robot Execution Failures	Multivariate, Time-Series	Classification	Integer	463	90	1999
Synthetic Control Chart Time Series	Time-Series	Classification, Clustering	Real	600	NA	1999
Syskill and Webert Web Page Ratings	Multivariate, Text	Classification	Categorical	332	5	1998
UNIX User Data	Text, Sequential	NA	NA	NA	NA	NA
Volcanoes on Venus - JARtool experiment	Image	Classification	NA	NA	NA	NA
Statlog (Australian Credit Approval)	Multivariate	Classification	Categorical, Integer, Real	690	14	NA
Statlog (German Credit Data)	Multivariate	Classification	Categorical, Integer	1000	20	1994
Statlog (Heart)	Multivariate	Classification	Categorical, Real	270	13	NA
Statlog (Landsat Satellite)	Multivariate	Classification	Integer	6435	36	1993
Statlog (Image Segmentation)	Multivariate	Classification	Real	2310	19	1990
Statlog (Shuttle)	Multivariate	Classification	Integer	58000	9	NA
Statlog (Vehicle Silhouettes)	Multivariate	Classification	Integer	946	18	NA
Connectionist Bench (Nettalk Corpus)	Multivariate	NA	Categorical	20008	4	NA
Connectionist Bench (Sonar, Mines vs. Rocks)	Multivariate	Classification	Real	208	60	NA
Connectionist Bench (Vowel Recognition - Deterding Data)	NA	Classification	Real	528	10	NA
Economic Sanctions	Domain-Theory	NA	NA	NA	NA	NA
Protein Data	NA	NA	NA	NA	NA	NA
Cloud	Multivariate	NA	Real	1024	10	1989
CalIt2 Building People Counts	Multivariate, Time-Series	NA	Categorical, Integer	10080	4	2006
Dodgers Loop Sensor	Multivariate, Time-Series	NA	Categorical, Integer	50400	3	2006
Poker Hand	Multivariate	Classification	Categorical, Integer	1025010	11	2007
MAGIC Gamma Telescope	Multivariate	Classification	Real	19020	11	2007
UJI Pen Characters	Multivariate, Sequential	Classification	Integer	1364	NA	2007
Mammographic Mass	Multivariate	Classification	Integer	961	6	2007
Forest Fires	Multivariate	Regression	Real	517	13	2008
Reuters Transcribed Subset	Text	Classification	NA	200	NA	2008
Bag of Words	Text	Clustering	Integer	8000000	100000	2008
Concrete Compressive Strength	Multivariate	Regression	Real	1030	9	2007
Hill-Valley	Sequential	Classification	Real	606	101	2008
Arcene	Multivariate	Classification	Real	900	10000	2008
Dexter	Multivariate	Classification	Integer	2600	20000	2008
Dorothea	Multivariate	Classification	Integer	1950	100000	2008
Gisette	Multivariate	Classification	Integer	13500	5000	2008
Madelon	Multivariate	Classification	Real	4400	500	2008
Ozone Level Detection	Multivariate, Sequential, Time-Series	Classification	Real	2536	73	2008
Abscisic Acid Signaling Network	Multivariate	Causal-Discovery	Integer	300	43	2008
Parkinsons	Multivariate	Classification	Real	197	23	2008
Character Trajectories	Time-Series	Classification, Clustering	Real	2858	3	2008
Blood Transfusion Service Center	Multivariate	Classification	Real	748	5	2008
UJI Pen Characters (Version 2)	Multivariate, Sequential	Classification	Integer	11640	NA	2009
Semeion Handwritten Digit	Multivariate	Classification	Integer	1593	256	2008
SECOM	Multivariate	Classification, Causal-Discovery	Real	1567	591	2008
Plants	Multivariate	Clustering	Categorical	22632	70	2008
Libras Movement	Multivariate, Sequential	Classification, Clustering	Real	360	91	2009
Concrete Slump Test	Multivariate	Regression	Real	103	10	2009
Communities and Crime	Multivariate	Regression	Real	1994	128	2009
Acute Inflammations	Multivariate	Classification	Categorical, Integer	120	6	2009
Wine Quality	Multivariate	Classification, Regression	Real	4898	12	2009
URL Reputation	Multivariate, Time-Series	Classification	Integer, Real	2396130	3231961	2009
p53 Mutants	Multivariate	Classification	Real	16772	5409	2010
Parkinsons Telemonitoring	Multivariate	Regression	Integer, Real	5875	26	2009
Demospongiae	Multivariate	Classification	Integer	503	NA	2010
Opinosis Opinion / Review	Text	NA	NA	51	NA	2010
Breast Tissue	Multivariate	Classification	Real	106	10	2010
Cardiotocography	Multivariate	Classification	Real	2126	23	2010
Wall-Following Robot Navigation Data	Multivariate, Sequential	Classification	Real	5456	24	2010
Spoken Arabic Digit	Multivariate, Time-Series	Classification	Real	8800	13	2010
Localization Data for Person Activity	Univariate, Sequential, Time-Series	Classification	Real	164860	8	2010
AutoUniv	Multivariate	Classification	Categorical, Integer, Real	NA	NA	2010
Steel Plates Faults	Multivariate	Classification	Integer, Real	1941	27	2010
MiniBooNE particle identification	Multivariate	Classification	Real	130065	50	2010
YearPredictionMSD	Multivariate	Regression	Real	515345	90	2011
PEMS-SF	Multivariate, Time-Series	Classification	Real	440	138672	2011
OpinRank Review Dataset	Text	NA	NA	NA	NA	2011
Relative location of CT slices on axial axis	Domain-Theory	Regression	Real	53500	386	2011
Online Handwritten Assamese Characters Dataset	Multivariate, Sequential	Classification	Integer	8235	NA	2011
PubChem Bioassay Data	Multivariate	Classification	Integer, Real	NA	NA	2011
Record Linkage Comparison Patterns	Multivariate	Classification	Real	5749132	12	2011
Communities and Crime Unnormalized	Multivariate	Regression	Real	2215	147	2011
Vertebral Column	Multivariate	Classification	Real	310	6	2011
EMG Physical Action Data Set	Time-Series	Classification	Real	10000	8	2011
Vicon Physical Action Data Set	Time-Series	Classification	Real	3000	27	2011
Amazon Commerce reviews set	Multivariate, Text, Domain-Theory	Classification	Real	1500	10000	2011
Amazon Access Samples	Time-Series, Domain-Theory	Regression, Clustering, Causal-Discovery	NA	30000	20000	2011
Reuter_50_50	Multivariate, Text, Domain-Theory	Classification, Clustering	Real	2500	10000	2011
Farm Ads	Text	Classification	NA	4143	54877	2011
DBWorld e-mails	Text	Classification	NA	64	4702	2011
KEGG Metabolic Relation Network (Directed)	Multivariate, Univariate, Text	Classification, Regression, Clustering	Integer, Real	53414	24	2011
KEGG Metabolic Reaction Network (Undirected)	Multivariate, Univariate, Text	Classification, Regression, Clustering	Integer, Real	65554	29	2011
Bank Marketing	Multivariate	Classification	Real	45211	17	2012
YouTube Comedy Slam Preference Data	Text	Classification	NA	1138562	3	2012
Gas Sensor Array Drift Dataset	Multivariate	Classification	Real	13910	128	2012
ILPD (Indian Liver Patient Dataset)	Multivariate	Classification	Integer, Real	583	10	2012
OPPORTUNITY Activity Recognition	Multivariate, Time-Series	Classification	Real	2551	242	2012
Nomao	Univariate	Classification	Real	34465	120	2012
SMS Spam Collection	Multivariate, Text, Domain-Theory	Classification, Clustering	Real	5574	NA	2012
Skin Segmentation	Univariate	Classification	Real	245057	4	2012
Planning Relax	Univariate	Classification	Real	182	13	2012
PAMAP2 Physical Activity Monitoring	Multivariate, Time-Series	Classification	Real	3850505	52	2012
Restaurant & consumer data	Multivariate	NA	NA	138	47	2012
CNAE-9	Multivariate, Text	Classification	Integer	1080	857	2012
Individual household electric power consumption	Multivariate, Time-Series	Regression, Clustering	Real	2075259	9	2012
seeds	Multivariate	Classification, Clustering	Real	210	7	2012
Northix	Multivariate, Univariate, Text	Classification	Integer, Real	115	200	2012
QtyT40I10D100K	Sequential	NA	Integer	3960456	4	2012
Legal Case Reports	Text	Classification	NA	NA	NA	2012
Human Activity Recognition Using Smartphones	Multivariate, Time-Series	Classification, Clustering	NA	10299	561	2012
One-hundred plant species leaves data set	NA	Classification	Real	1600	64	2012
Energy efficiency	Multivariate	Classification, Regression	Integer, Real	768	8	2012
Yacht Hydrodynamics	Multivariate	Regression	Real	308	7	2013
Fertility	Multivariate	Classification, Regression	Real	100	10	2013
Daphnet Freezing of Gait	Multivariate, Time-Series	Classification	Real	237	9	2013
3D Road Network (North Jutland, Denmark)	Sequential, Text	Regression, Clustering	Real	434874	4	2013
ISTANBUL STOCK EXCHANGE	Multivariate, Univariate, Time-Series	Classification, Regression	Real	536	8	2013
Buzz in social media	Time-Series, Multivariate	Regression, Classification	Integer, Real	140000	77	2013
First-order theorem proving	Multivariate	Classification	Real	6118	51	2013
Wearable Computing: Classification of Body Postures and Movements (PUC-Rio)	Sequential	Classification	Integer, Real	165632	18	2013
Gas sensor arrays in open sampling settings	Multivariate, Time-Series	Classification	Real	18000	1950000	2013
Climate Model Simulation Crashes	Multivariate	Classification	Real	540	18	2013
MicroMass	Multivariate	Classification	Real	931	1300	2013
QSAR biodegradation	Multivariate	Classification	Integer, Real	1055	41	2013
BLOGGER	Multivariate	Classification	NA	100	6	2013
Daily and Sports Activities	Multivariate, Time-Series	Classification, Clustering	Real	9120	5625	2013
User Knowledge Modeling	Multivariate	Classification, Clustering	Integer	403	5	2013
Reuters RCV1 RCV2 Multilingual, Multiview Text Categorization Test collection	Multivariate	Classification	Real	111740	NA	2013
NYSK	Multivariate, Sequential, Text	Clustering	NA	10421	7	2013
Turkiye Student Evaluation	Multivariate	Classification, Clustering	NA	5820	33	2013
ser Knowledge Modeling Data (Students’ Knowledge Levels on DC Electrical Machines)	Multivariate	Classification	Real	403	5	2013
EEG Eye State	Multivariate, Sequential, Time-Series	Classification	Integer, Real	14980	15	2013
Physicochemical Properties of Protein Tertiary Structure	Multivariate	Regression	Real	45730	9	2013
seismic-bumps	Multivariate	Classification	Real	2584	19	2013
banknote authentication	Multivariate	Classification	Real	1372	5	2013
USPTO Algorithm Challenge, run by NASA-Harvard Tournament Lab and TopCoder Problem: Pat	Domain-Theory	Classification	Integer	306	5	2013
YouTube Multiview Video Games Dataset	Multivariate, Text	Classification, Clustering	Integer, Real	120000	1000000	2013
Gas Sensor Array Drift Dataset at Different Concentrations	Multivariate, Time-Series	Classification, Regression, Clustering, Causa	Real	13910	129	2013
Activities of Daily Living (ADLs) Recognition Using Binary Sensors	Multivariate, Sequential, Time-Series	Classification, Clustering	NA	2747	NA	2013
SkillCraft1 Master Table Dataset	Multivariate	Regression	Integer, Real	3395	20	2013
Weight Lifting Exercises monitored with Inertial Measurement Units	Multivariate	Classification	Real	39242	152	2013
SML2010	Multivariate, Sequential, Time-Series, Text	Regression	Real	4137	24	2014
Bike Sharing Dataset	Univariate	Regression	Integer, Real	17389	16	2013
Predict keywords activities in a online social media	Multivariate, Sequential, Time-Series	NA	Integer, Real	51	35	2013
Thoracic Surgery Data	Multivariate	Classification	Integer, Real	470	17	2013
EMG dataset in Lower Limb	Multivariate, Time-Series	NA	Real	132	5	2014
SUSY	NA	Classification	Real	5000000	18	2014
HIGGS	NA	Classification	Real	11000000	28	2014
Qualitative_Bankruptcy	Multivariate	Classification	NA	250	7	2014
LSVT Voice Rehabilitation	Multivariate	Classification	Real	126	309	2014
Dataset for ADL Recognition with Wrist-worn Accelerometer	Multivariate, Time-Series	Classification, Clustering	NA	NA	3	2014
Wilt	Multivariate	Classification	NA	4889	6	2014
User Identification From Walking Activity	Univariate, Sequential, Time-Series	Classification, Clustering	Real	NA	NA	2014
Activity Recognition from Single Chest-Mounted Accelerometer	Univariate, Sequential, Time-Series	Classification, Clustering	Real	NA	NA	2014
Leaf	Multivariate	Classification	Real	340	16	2014
Dresses_Attribute_Sales	Text	Classification, Clustering	NA	501	13	2014
Tamilnadu Electricity Board Hourly Readings	Multivariate	Classification, Regression, Clustering	Real	45781	5	2013
Airfoil Self-Noise	Multivariate	Regression	Real	1503	6	2014
Wholesale customers	Multivariate	Classification, Clustering	Integer	440	8	2014
Twitter Data set for Arabic Sentiment Analysis	Text	Classification	NA	2000	2	2014
Combined Cycle Power Plant	Multivariate	Regression	Real	9568	4	2014
Urban Land Cover	Multivariate	Classification	NA	168	148	2014
Diabetes 130-US hospitals for years 1999-2008	Multivariate	Classification, Clustering	Integer	100000	55	2014
Bach Choral Harmony	Sequential	Classification	NA	5665	17	2014
StoneFlakes	Multivariate	Classification, Clustering, Causal-Discovery	Real	79	8	2014
Tennis Major Tournament Match Statistics	Multivariate	Classification, Regression, Clustering	Integer, Real	127	42	2014
Parkinson Speech Dataset with Multiple Types of Sound Recordings	Multivariate	Classification, Regression	Integer, Real	1040	26	2014
Gesture Phase Segmentation	Multivariate, Sequential, Time-Series	Classification, Clustering	Real	9900	50	2014
Perfume Data	Univariate, Domain-Theory	Classification, Clustering	Integer	560	2	2014
BlogFeedback	Multivariate	Regression	Integer, Real	60021	281	2014
REALDISP Activity Recognition Dataset	Multivariate, Time-Series	Classification	Real	1419	120	2014
Newspaper and magazine images segmentation dataset	NA	Classification	NA	101	NA	2014
AAAI 2014 Accepted Papers	Multivariate	Clustering	NA	399	6	2014
Gas sensor array under flow modulation	Multivariate, Time-Series	Classification, Regression	Real	58	120432	2014
Gas sensor array exposed to turbulent gas mixtures	Multivariate, Time-Series	Classification, Regression	Real	180	150000	2014
UJIIndoorLoc	Multivariate	Classification, Regression	Integer, Real	21048	529	2014
Sentence Classification	Text	Classification	Integer	NA	NA	2014
Dow Jones Index	Time-Series	Classification, Clustering	Integer, Real	750	16	2014
sEMG for Basic Hand movements	Time-Series	Classification	Real	3000	2500	2014
AAAI 2013 Accepted Papers	Multivariate	Clustering	NA	150	5	2014
Geographical Original of Music	Multivariate	Classification, Regression	Real	1059	68	2014
Condition Based Maintenance of Naval Propulsion Plants	Multivariate	Regression	Real	11934	16	2014
Grammatical Facial Expressions	Multivariate, Sequential	Classification, Clustering	Real	27965	100	2014
NoisyOffice	Multivariate	Classification, Regression	Real	216	216	2015
MHEALTH Dataset	Multivariate, Time-Series	Classification	Real	120	23	2014
Student Performance	Multivariate	Classification, Regression	Integer	649	33	2014
ElectricityLoadDiagrams20112014	Time-Series	Regression, Clustering	Real	370	140256	2015
Gas sensor array under dynamic gas mixtures	Multivariate, Time-Series	Classification, Regression	Real	4178504	19	2015
microblogPCU	Multivariate, Univariate, Sequential, Text	Classification, Causal-Discovery	Integer, Real	221579	20	2015
Firm-Teacher_Clave-Direction_Classification	Multivariate	Classification	NA	10800	20	2015
Dataset for Sensorless Drive Diagnosis	Multivariate	Classification	Real	58509	49	2015
TV News Channel Commercial Detection Dataset	Multivariate	Classification, Clustering	Real	129685	12	2015
Phishing Websites	NA	Classification	Integer	2456	30	2015
Greenhouse Gas Observing Network	Multivariate, Time-Series	Regression	Real	2921	5232	2015
Diabetic Retinopathy Debrecen Data Set	Multivariate	Classification	Integer, Real	1151	20	2014
HIV-1 protease cleavage	Multivariate	Classification	Categorical	6590	1	2015
Sentiment Labelled Sentences	Text	Classification	NA	3000	NA	2015
Online News Popularity	Multivariate	Classification, Regression	Integer, Real	39797	61	2015
Forest type mapping	Multivariate	Classification	NA	326	27	2015
wiki4HE	Multivariate	Regression, Clustering, Causal-Discovery	NA	913	53	2015
Online Video Characteristics and Transcoding Time Dataset	Multivariate	Regression	Integer, Real	168286	11	2015
Chronic_Kidney_Disease	Multivariate	Classification	Real	400	25	2015
Machine Learning based ZZAlpha Ltd. Stock Recommendations 2012-2014	Sequential, Time-Series	Classification	Real	314080	0	2015
Folio	Multivariate	Classification, Clustering	NA	637	20	2015
Taxi Service Trajectory - Prediction Challenge, ECML PKDD 2015	Multivariate, Sequential, Time-Series, Domain-Theory	Clustering, Causal-Discovery	Real	1710671	9	2015
Cuff-Less Blood Pressure Estimation	Multivariate	Classification, Regression	Real	12000	3	2015
Smartphone-Based Recognition of Human Activities and Postural Transitions	Multivariate, Time-Series	Classification	Real	10929	561	2015
Mice Protein Expression	Multivariate	Classification, Clustering	Real	1080	82	2015
UJIIndoorLoc-Mag	Multivariate, Sequential, Time-Series	Classification, Regression, Clustering	Integer, Real	40000	13	2015
Heterogeneity Activity Recognition	Multivariate, Time-Series	Classification, Clustering	Real	43930257	16	2015
Educational Process Mining (EPM): A Learning Analytics Data Set	Multivariate, Sequential, Time-Series	Classification, Regression, Clustering	Integer	230318	13	2015
HEPMASS	Multivariate	Classification	Real	10500000	28	2016
Indoor User Movement Prediction from RSS data	Multivariate, Sequential, Time-Series	Classification	Real	13197	4	2016
Open University Learning Analytics dataset	Multivariate, Sequential, Time-Series	Classification, Regression, Clustering	Integer	NA	NA	2015
default of credit card clients	Multivariate	Classification	Integer, Real	30000	24	2016
Mesotheliomaâ<U+0080><U+0099>s disease data set	Multivariate	Classification	Real	324	34	2016
Online Retail	Multivariate, Sequential, Time-Series	Classification, Clustering	Integer, Real	541909	8	2015
SIFT10M	Multivariate	Causal-Discovery	Integer	11164866	128	2016
GPS Trajectories	Multivariate	Classification, Regression	Real	163	15	2016
Detect Malacious Executable(AntiVirus)	Multivariate	Classification	Real	373	513	2016
Occupancy Detection	Multivariate, Time-Series	Classification	Real	20560	7	2016
Improved Spiral Test Using Digitized Graphics Tablet for Monitoring Parkinsonâ<U+0080><U+0099>s Disease	Multivariate	Classification, Regression, Clustering	Real	40	7	2016
News Aggregator	Multivariate	Classification, Clustering	NA	422937	5	2016
Air Quality	Multivariate, Time-Series	Regression	Real	9358	15	2016
Twin gas sensor arrays	Multivariate, Time-Series, Domain-Theory	Classification, Regression	Real	640	480000	2016
Gas sensors for home activity monitoring	Multivariate, Time-Series	Classification	Real	919438	11	2016
Facebook Comment Volume Dataset	Multivariate	Regression	Integer, Real	40949	54	2016
Smartphone Dataset for Human Activity Recognition (HAR) in Ambient Assisted Living (AAL)	Time-Series	Classification	Real	5744	561	2016
Polish companies bankruptcy data	Multivariate	Classification	Real	10503	64	2016
Activity Recognition system based on Multisensor data fusion (AReM)	Multivariate, Sequential, Time-Series	Classification	Real	42240	6	2016
Dota2 Games Results	Multivariate	Classification	NA	102944	116	2016
Facebook metrics	Multivariate	Regression	Integer	500	19	2016
UbiqLog (smartphone lifelogging)	Multivariate	Causal-Discovery	NA	9782222	NA	2016
NIPS Conference Papers 1987-2015	Text	Clustering	Integer	11463	5812	2016
HTRU2	Multivariate	Classification, Clustering	Real	17898	9	2017
Drug consumption (quantified)	Multivariate	Classification	Real	1885	32	2016
Appliances energy prediction	Multivariate, Time-Series	Regression	Real	19735	29	2017
Miskolc IIS Hybrid IPS	Text	Classification, Clustering, Causal-Discovery	Integer	1540	67	2016
KDC-4007 dataset Collection	Multivariate, Text	Classification, Regression	Integer	4007	NA	2017
Geo-Magnetic field and WLAN dataset for indoor localisation from wristband and smartphone	Multivariate, Sequential, Time-Series	Classification, Regression, Clustering	Integer, Real	153540	25	2017
DrivFace	Multivariate	Classification, Regression, Clustering	Real	606	6400	2016
Website Phishing	Multivariate	Classification	Integer	1353	10	2016
YouTube Spam Collection	Text	Classification	NA	1956	5	2017
Beijing PM2.5 Data	Multivariate, Time-Series	Regression	Integer, Real	43824	13	2017
Cargo 2000 Freight Tracking and Tracing	Multivariate, Sequential	Classification, Regression	Integer	3942	98	2016
Cervical cancer (Risk Factors)	Multivariate	Classification	Integer, Real	858	36	2017
Quality Assessment of Digital Colposcopies	Multivariate	Classification	Real	287	69	2017
KASANDR	Multivariate	Causal-Discovery	Integer	17764280	2158859	2017
FMA: A Dataset For Music Analysis	Multivariate, Time-Series	Classification, Clustering	Real	106574	518	2017
Air quality	Multivariate, Time-Series	Regression	Real	9358	15	2016
Epileptic Seizure Recognition	Multivariate, Time-Series	Classification, Clustering	Integer, Real	11500	179	2017
Devanagari Handwritten Character Dataset	NA	Classification	Integer	92000	NA	2016
Stock portfolio performance	Multivariate	Regression	Real	315	12	2016
MoCap Hand Postures	Multivariate	Classification, Clustering	Integer, Real	78095	38	2016
Early biomarkers of Parkinson<U+0092>s disease based on natural connected speech	Multivariate	Classification, Regression	Integer, Real	130	65	2017
Data for Software Engineering Teamwork Assessment in Education Setting	Sequential, Time-Series	Classification	Integer, Real	74	102	2017
PM2.5 Data of Five Chinese Cities	Multivariate, Time-Series	Regression	Integer, Real	52854	86	2017
Parkinson Disease Spiral Drawings Using Digitized Graphics Tablet	Multivariate	Classification, Regression, Clustering	Integer	77	7	2017
Sales_Transactions_Dataset_Weekly	Multivariate, Time-Series	Clustering	Integer, Real	811	53	2017
Las Vegas Strip	NA	Classification, Regression	Integer	504	20	2017
Eco-hotel	Text	NA	NA	401	1	2017
MEU-Mobile KSD	Multivariate	Classification	Integer, Real	2856	71	2016
Crowdsourced Mapping	Multivariate	Classification	NA	10546	29	2016
gene expression cancer RNA-Seq	Multivariate	Classification, Clustering	Real	801	20531	2016
Hybrid Indoor Positioning Dataset from WiFi RSSI, Bluetooth and magnetometer	Multivariate, Sequential, Time-Series	Classification	Real	1540	65	2016
chestnut â<U+0080><U+0093> LARVIC	NA	Classification, Clustering	NA	1451	3	2017
Burst Header Packet (BHP) flooding attack on Optical Burst Switching (OBS) Network	Text	Classification	Integer	1075	22	2017
Motion Capture Hand Postures	Multivariate	Classification, Clustering	Real	78095	38	2017
Anuran Calls (MFCCs)	Multivariate	Classification, Clustering	Real	7195	22	2017
TTC-3600: Benchmark dataset for Turkish text categorization	Text	Classification, Clustering	Integer	3600	4814	2017
Gastrointestinal Lesions in Regular Colonoscopy	Multivariate	Classification	Real	76	698	2016
Daily Demand Forecasting Orders	Time-Series	Regression	Integer	60	13	2017
Paper Reviews	Text	Classification, Regression	Integer	405	10	2017
extention of Z-Alizadeh sani dataset	NA	Classification	Integer, Real	303	59	2017
Z-Alizadeh Sani	NA	Classification	Integer, Real	303	56	2017
Dynamic Features of VirusShare Executables	Multivariate, Time-Series	Classification, Regression	Integer	107888	482	2017
IDA2016Challenge	Multivariate	Classification	Integer	76000	171	2017
DSRC Vehicle Communications	Sequential, Text	Clustering	Real	10000	5	2017
Mturk User-Perceived Clusters over Images	Multivariate, Text	Clustering	Integer	180	500	2016
Character Font Images	Multivariate	Classification	Integer, Real	745000	411	2016
DeliciousMIL: A Data Set for Multi-Label Multi-Instance Learning with Instance Labels	Text	Classification	Integer	12234	8519	2016
Autistic Spectrum Disorder Screening Data for Children	Multivariate	Classification	Integer	292	21	2017
Autistic Spectrum Disorder Screening Data for Adolescent	Multivariate	Classification	Integer	104	21	2017
APS Failure at Scania Trucks	Multivariate	Classification	Integer, Real	60000	171	2017
Wireless Indoor Localization	Multivariate	Classification	Real	2000	7	2017
HCC Survival	Multivariate	Classification	Integer, Real	165	49	2017
CSM (Conventional and Social Media Movies) Dataset 2014 and 2015	Multivariate	Classification, Regression	Integer	217	12	2017
University of Tehran Question Dataset 2016 (UTQD.2016)	Text	Classification	NA	1175	3	2017
Autism Screening Adult	NA	Classification	Integer	704	21	2017
Activity recognition with healthy older people using a batteryless wearable sensor	Sequential	Classification	Real	75128	9	2016
Immunotherapy Dataset	Univariate	Classification	Integer, Real	90	8	2018
Cryotherapy Dataset	Univariate	Classification	Integer, Real	90	7	2018
OCT data & Color Fundus Images of Left & Right Eyes	Multivariate	Classification	Real	50	2	2016
Discrete Tone Image Dataset	Multivariate	Classification	NA	71	11	2018
News Popularity in Multiple Social Media Platforms	Multivariate, Time-Series, Text	Regression	Integer, Real	93239	11	2018
Ultrasonic flowmeter diagnostics	Multivariate	Classification	Real	540	173	2018
ICMLA 2014 Accepted Papers Data Set	Multivariate	Classification, Clustering	NA	105	5	2018
BLE RSSI Dataset for Indoor localization and Navigation	Multivariate, Sequential, Time-Series	Classification, Clustering	Integer	6611	15	2018
Container Crane Controller Data Set	Univariate, Domain-Theory	Classification, Regression	Real	15	3	2018
Residential Building Data Set	Multivariate	Regression	Real	372	105	2018
Health News in Twitter	Text	Clustering	Real	58000	25000	2018
chipseq	Sequential	Classification	Integer	4960	NA	2018
SGEMM GPU kernel performance	Multivariate	Regression	Integer	241600	18	2018
Repeat Consumption Matrices	Multivariate	Clustering	Real	130000	21000	2018
detection_of_IoT_botnet_attacks_N_BaIoT	Multivariate, Sequential	Classification, Clustering	Real	7062606	115	2018
Absenteeism at work	Multivariate, Time-Series	Classification, Clustering	Integer, Real	740	21	2018
SCADI	Multivariate	Classification, Clustering	NA	70	206	2018
Condition monitoring of hydraulic systems	Multivariate, Time-Series	Classification, Regression	Real	2205	43680	2018
Carbon Nanotubes	Univariate	Regression	Real	10721	8	2018
Optical Interconnection Network	Multivariate	Classification, Regression	Integer, Real	640	10	2018
Sports articles for objectivity analysis	Multivariate, Text	Classification	Integer	1000	59	2018
Breast Cancer Coimbra	Multivariate	Classification	Integer	116	10	2018
GNFUV Unmanned Surface Vehicles Sensor Data	Multivariate, Time-Series	Regression	Real	1672	5	2018
Dishonest Internet users Dataset	Multivariate	Classification, Clustering	NA	322	5	2018
Victorian Era Authorship Attribution	Text	Classification	NA	93600	1000	2018
Simulated Falls and Daily Living Activities Data Set	Time-Series	Classification	Integer	3060	138	2018
Multimodal Damage Identification for Humanitarian Computing	Multivariate, Text	Classification	Integer	5879	NA	2018
EEG Steady-State Visual Evoked Potential Signals	Multivariate, Time-Series	Classification, Regression	Integer	9200	16	2018
Roman Urdu Data Set	Text	Classification	NA	20000	2	2018
Avila	Multivariate	Classification	Real	20867	10	2018
PANDOR	Multivariate	Recommendation	Categorical	NA	NA	2018
Drug Review Dataset (Druglib.com)	Multivariate, Text	Classification, Regression, Clustering	Integer	4143	8	2018
Drug Review Dataset (Drugs.com)	Multivariate, Text	Classification, Regression, Clustering	Integer	215063	6	2018
Physical Unclonable Functions	Multivariate	Classification	Integer	6000000	129	2018
Superconductivty Data	Multivariate	Regression	Real	21263	81	2018
WESAD (Wearable Stress and Affect Detection)	Multivariate, Time-Series	Classification, Regression	Real	63000000	12	2018
GNFUV Unmanned Surface Vehicles Sensor Data Set 2	Multivariate, Sequential, Time-Series	Regression	Real	10190	6	2018
Student Academics Performance	Multivariate	Classification	NA	300	22	2018
Online Shoppers Purchasing Intention Dataset	Multivariate	Classification, Clustering	Integer, Real	12330	18	2018
PMU-UD	Univariate	Classification	NA	5180	9	2018
Parkinson’s Disease Classification	Multivariate	Classification	Integer, Real	756	754	2018
Electrical Grid Stability Simulated Data	Multivariate	Classification, Regression	Real	10000	14	2018
Caesarian Section Classification Dataset	Univariate	Classification	Integer	80	5	2018
BAUM-1	Time-Series	Classification	NA	1184	NA	2018
BAUM-2	Time-Series	Classification	NA	1047	NA	2018
Audit Data	Multivariate	Classification	Real	777	18	2018
BuddyMove Data Set	Multivariate, Text	Classification, Clustering	Real	249	7	2018
Real estate valuation data set	Multivariate	Regression	Integer, Real	414	7	2018
Early biomarkers of Parkinsonâ<U+0080><U+0099>s disease based on natural connected speech Data Set	Multivariate	Classification	Real	NA	NA	2018
Somerville Happiness Survey	NA	Classification	Integer	143	7	2018
2.4 GHZ Indoor Channel Measurements	Multivariate	Classification	Real	7840	5	2018
EMG data for gestures	Time-Series	Classification	Real	30000	6	2019
Parking Birmingham	Multivariate, Univariate, Sequential, Time-Series	Classification, Regression, Clustering	Real	35717	4	2019
Behavior of the urban traffic of the city of Sao Paulo in Brazil	Multivariate, Time-Series	Classification, Regression	Integer, Real	135	18	2018
Travel Reviews	Multivariate, Text	Classification, Clustering	Real	980	11	2018
Tarvel Review Ratings	Multivariate, Text	Classification, Clustering	Real	5456	25	2018
Rice Leaf Diseases	Multivariate	Classification	Integer	120	NA	2019
Gas sensor array temperature modulation	Multivariate, Time-Series	Classification, Regression	Real	4095000	20	2019
Facebook Live Sellers in Thailand	Multivariate	Clustering	Integer	7051	12	2019
Parkinson Dataset with replicated acoustic features	Multivariate	Classification	NA	240	46	2019
Metro Interstate Traffic Volume	Multivariate, Sequential, Time-Series	Regression	Integer, Real	48204	9	2019
Query Analytics Workloads Dataset	Multivariate	Regression, Clustering	Real	260000	8	2019
Wave Energy Converters	Multivariate	Regression	Real	288000	49	2019
PPG-DaLiA	Multivariate, Time-Series	Regression	Real	8300000	11	2019
Alcohol QCM Sensor Dataset	Multivariate	Classification, Regression, Clustering	Real	125	8	2019
Divorce Predictors data set	Multivariate, Univariate	Classification	Integer	170	54	2019
Incident management process enriched event log	Multivariate, Sequential	Regression, Clustering	Integer	141712	36	2019
Opinion Corpus for Lebanese Arabic Reviews (OCLAR)	Text	Classification	Integer	3916	3916	2019
MEx	Time-Series	Classification, Clustering	Real	6262	710	2019
Beijing Multi-Site Air-Quality Data	Multivariate, Time-Series	Regression	Integer, Real	420768	18	2019
Online Retail II	Multivariate, Sequential, Time-Series, Text	Classification, Regression, Clustering	Integer, Real	1067371	8	2019
Hepatitis C Virus (HCV) for Egyptian patients	Multivariate	Classification	Integer, Real	1385	29	2019
QSAR fish toxicity	Multivariate	Regression	Real	908	7	2019
QSAR aquatic toxicity	Multivariate	Regression	Real	546	9	2019
Human Activity Recognition from Continuous Ambient Sensor Data	Multivariate, Sequential, Time-Series	Classification	Integer, Real	13956534	37	2019
WISDM Smartphone and Smartwatch Activity and Biometrics Dataset	Multivariate, Time-Series	Classification	Real	15630426	6	2019
QSAR oral toxicity	Multivariate	Classification	NA	8992	1024	2019
QSAR androgen receptor	Multivariate	Classification	NA	1687	1024	2019
QSAR Bioconcentration classes dataset	Multivariate	Classification, Regression	NA	779	14	2019
QSAR fish bioconcentration factor (BCF)	Multivariate	Regression	Integer, Real	1056	7	2019
A study of Asian Religious and Biblical Texts	Multivariate, Text	Classification, Clustering	Integer	590	8265	2019
Real-time Election Results: Portugal 2019	Multivariate, Time-Series, Text	Regression	Integer, Real	21643	29	2019
Bias correction of numerical prediction model temperature forecast	Multivariate	Regression	Real	7750	25	2020
Bar Crawl: Detecting Heavy Drinking	Multivariate, Time-Series	Classification, Regression	Real	14057567	3	2020
Kitsune Network Attack Dataset	Multivariate, Sequential, Time-Series	Classification, Clustering, Causal-Discovery	Real	27170754	115	2019
Shoulder Implant X-Ray Manufacturer Classification	Multivariate	Classification	Real	597	1	2020
Speaker Accent Recognition	Multivariate	Classification	Real	329	12	2020
Heart failure clinical records	Multivariate	Classification, Regression, Clustering	Integer, Real	299	13	2020
Deepfakes: Medical Image Tamper Detection	Multivariate	Classification	Real	20000	200000	2020
selfBACK	Time-Series	Classification, Clustering	Real	26136	6	2020
South German Credit	Multivariate	Classification, Regression, Clustering	Integer, Real	1000	21	2019
Exasens	Multivariate	Classification, Clustering	Integer	399	4	2020
Swarm Behaviour	Multivariate	Classification	Real	24017	2400	2020
Crop mapping using fused optical-radar data set	Multivariate, Time-Series	Classification	Real	325834	175	2020
BitcoinHeistRansomwareAddressDataset	Multivariate, Time-Series	Classification, Clustering	Integer, Real	2916697	10	2020
Facebook Large Page-Page Network	Multivariate	Classification	NA	22470	4714	2020
Amphibians	Multivariate	Classification	Integer, Real	189	23	2020
Early stage diabetes risk prediction dataset.	Multivariate	Classification	NA	520	17	2020
Turkish Spam V01	Text	Classification	NA	826	2	2019
Stock keeping units	Multivariate	Clustering	Integer, Real	2279	9	2019
Demand Forecasting for a store	Multivariate	Regression	Integer	28764	8	2019
Detect Malware Types	Multivariate, Time-Series, Text	Classification	NA	7107	280	2019
Wave Energy Converters	Multivariate	Regression	Real	288000	49	2019
Youtube cookery channels viewers comments in Hinglish	Multivariate, Text	Classification	NA	9800	3	2019
Pedestrian in Traffic Dataset	Multivariate, Sequential, Time-Series	Classification, Regression, Causal-Discovery	Real	4760	14	2019
Cervical Cancer Behavior Risk	Multivariate, Univariate	Classification, Clustering	Integer	72	19	2019
Sattriya_Dance_Single_Hand_Gestures Dataset	Multivariate	Classification	NA	1450	NA	2019
Divorce Predictors data set	Multivariate, Univariate	Classification	Integer	170	54	2019
3W dataset	Multivariate, Time-Series	Classification, Clustering	Integer, Real	1984	8	2019
Malware static and dynamic features VxHeaven and Virus Total	Multivariate	Classification	Integer, Real	2955	1087	2019
Internet Firewall Data	Multivariate	Classification	NA	65532	12	2019
User Profiling and Abusive Language Detection Dataset	NA	Classification	NA	65919	3	2019
Estimation of obesity levels based on eating habits and physical condition	Multivariate	Classification, Regression, Clustering	Integer	2111	17	2019
Rice (Cammeo and Osmancik)	Multivariate	Classification	Real	3810	8	2019
Vehicle routing and scheduling problems	Multivariate	Clustering	Integer, Real	18	9	2019
Algerian Forest Fires Dataset	Multivariate	Classification, Regression	Real	244	12	2019
Breath Metabolomics	Multivariate, Time-Series	Classification, Clustering	Real	104	1656	2019
Horton General Hospital	Multivariate, Time-Series	Causal-Discovery	Integer	139	6	2019
UrbanGB, urban road accidents coordinates labelled by the urban center	Univariate	Clustering	Real	360177	2	2019
Gas Turbine CO and NOx Emission Data Set	Multivariate	Regression, Clustering	Real	36733	11	2019
Activity recognition using wearable physiological measurements	Multivariate	Classification	Real	4480	533	2019
clickstream data for online shopping	Multivariate, Sequential	Classification, Regression, Clustering	Integer, Real	165474	14	2019
CNNpred: CNN-based stock market prediction using a diverse set of variables	Sequential, Time-Series	Classification, Regression	Real	1985	84	2019
Apartment for rent classified	Multivariate	Classification, Regression, Clustering	NA	10000	22	2019
: Simulated Data set of Iraqi tourism places	Multivariate	Classification, Clustering	NA	232	16	2020
Nasarian CAD Dataset	Multivariate	Classification	NA	150	52	2020
Monolithic Columns in Troad and Mysia Region	Multivariate	Classification	Real	11	19	2020
Bar Crawl: Detecting Heavy Drinking	Multivariate, Time-Series	Classification, Regression	Real	14057567	3	2020
Seoul Bike Sharing Demand	Multivariate	Regression	Integer, Real	8760	14	2020
Person Classification Gait Data	Multivariate	Classification	Real	48	321	2020
Shill Bidding Dataset	Multivariate	Classification, Clustering	NA	6321	13	2020
Iranian Churn Dataset	Multivariate	Classification, Regression	Integer	3150	13	2020
Unmanned Aerial Vehicle (UAV) Intrusion Detection	Multivariate	Classification	Real	17256	55	2020
Bone marrow transplant: children	Multivariate	Classification, Regression	Integer, Real	187	39	2020
Exasens	Multivariate	Classification, Clustering	Integer	399	4	2020
COVID-19 Surveillance	Multivariate	Classification	NA	14	7	2020
Refractive errors	Multivariate	Classification	Integer	467	79	2020
Shoulder Implant X-Ray Manufacturer Classification	Multivariate	Classification	Real	597	1	2020
CLINC150	Text	Classification	NA	23700	NA	2020
HCV data	Multivariate	Classification, Clustering	Integer, Real	615	14	2020
Taiwanese Bankruptcy Prediction	Multivariate	Classification	Integer	6819	96	2020
South German Credit (UPDATE)	Multivariate	Classification, Regression, Clustering	Integer, Real	1000	21	2020
IIWA14-R820-Gazebo-Dataset-10Trajectories	NA	Regression	Integer	NA	NA	2020
Guitar Chords finger positions	Text	Classification	NA	2633	5	2020
Russian Corpus of Biographical Texts	Text	Classification	NA	200	2	2020
Codon usage	Multivariate	Classification, Clustering	NA	13028	69	2020
Intelligent Media Accelerometer and Gyroscope (IM-AccGyro) Dataset	Time-Series	Classification	Real	800	9	2020
Myocardial infarction complications	Multivariate	Classification	Real	1700	124	2020
Hungarian Chickenpox Cases	Time-Series	Regression	Real	521	20	2021
Simulated data for survival modelling	Multivariate, Time-Series	Regression	Integer, Real	120000	25	2018
Student Performance on an entrance examination	Multivariate	Classification	NA	666	11	2018
Chemical Composition of Ceramic Samples	Multivariate	Classification, Clustering	Real	88	19	2019
Labeled Text Forum Threads Dataset	Text	Classification	Integer	200	9	2019
Stock keeping units	Multivariate	Clustering	Integer, Real	2279	9	2019
BLE RSSI dataset for Indoor localization	Sequential, Time-Series	Classification	Integer	23570	5	2019
Basketball dataset	Time-Series	Classification	Integer	10000	7	2019
GitHub MUSAE	Multivariate	Classification	NA	37700	4006	2019
Anticancer peptides	Sequential	Classification	NA	1850	2	2019
Monolithic Columns in Troad and Mysia Region	Multivariate	Classification	Real	11	19	2020
Gender by Name	Text	Classification, Clustering	NA	147270	4	2020
Iranian Churn Dataset	Multivariate	Classification, Regression	Integer	3150	13	2020
Unmanned Aerial Vehicle (UAV) Intrusion Detection	Multivariate	Classification	Real	17256	55	2020
Shoulder Implant Manufacture Classification	Multivariate	Classification	NA	597	1	2020
LastFM Asia Social Network	Multivariate	Classification	NA	7624	7842	2020
Wheat kernels	Multivariate	Classification	Real	314	15	2020
Productivity Prediction of Garment Employees	Multivariate, Time-Series	Classification, Regression	Integer, Real	1197	15	2020
Multi-view Brain Networks	Multivariate	Classification, Clustering	Integer	70	70	2020
LastFM Asia Social Network	Multivariate	Classification	NA	7624	7842	2020
Wisesight Sentiment Corpus	Multivariate, Text	Classification	NA	26737	4	2020
AI4I 2020 Predictive Maintenance Dataset	Multivariate, Time-Series	Classification, Regression, Causal-Discovery	Real	10000	14	2020
Dry Bean Dataset	Multivariate	Classification	Integer, Real	13611	17	2020
in-vehicle coupon recommendation	Multivariate	Classification	NA	12684	23	2020
Gait Classification	Multivariate	Classification	Real	48	321	2020
Wikipedia Math Essentials	Time-Series	Regression	Real	731	1068	2021
Wikipedia Math Essentials	Time-Series	Regression	Real	731	1068	2021
Synchronous Machine Data Set	Multivariate	Regression	Real	557	5	2021

auto_mpg<-ucimlr::auto_mpg

knitr::kable(head(auto_mpg,100))%>%
  kableExtra::kable_styling("striped")%>%
  kableExtra::scroll_box(width = "100%",height="300px")

mpg	cylinders	displacement	horsepower	weight	acceleration	model_year	origin	car_name
18	8	307.0	130	3504	12.0	70	1	chevrolet chevelle malibu
15	8	350.0	165	3693	11.5	70	1	buick skylark 320
18	8	318.0	150	3436	11.0	70	1	plymouth satellite
16	8	304.0	150	3433	12.0	70	1	amc rebel sst
17	8	302.0	140	3449	10.5	70	1	ford torino
15	8	429.0	198	4341	10.0	70	1	ford galaxie 500
14	8	454.0	220	4354	9.0	70	1	chevrolet impala
14	8	440.0	215	4312	8.5	70	1	plymouth fury iii
14	8	455.0	225	4425	10.0	70	1	pontiac catalina
15	8	390.0	190	3850	8.5	70	1	amc ambassador dpl
15	8	383.0	170	3563	10.0	70	1	dodge challenger se
14	8	340.0	160	3609	8.0	70	1	plymouth ’cuda 340
15	8	400.0	150	3761	9.5	70	1	chevrolet monte carlo
14	8	455.0	225	3086	10.0	70	1	buick estate wagon (sw)
24	4	113.0	95	2372	15.0	70	3	toyota corona mark ii
22	6	198.0	95	2833	15.5	70	1	plymouth duster
18	6	199.0	97	2774	15.5	70	1	amc hornet
21	6	200.0	85	2587	16.0	70	1	ford maverick
27	4	97.0	88	2130	14.5	70	3	datsun pl510
26	4	97.0	46	1835	20.5	70	2	volkswagen 1131 deluxe sedan
25	4	110.0	87	2672	17.5	70	2	peugeot 504
24	4	107.0	90	2430	14.5	70	2	audi 100 ls
25	4	104.0	95	2375	17.5	70	2	saab 99e
26	4	121.0	113	2234	12.5	70	2	bmw 2002
21	6	199.0	90	2648	15.0	70	1	amc gremlin
10	8	360.0	215	4615	14.0	70	1	ford f250
10	8	307.0	200	4376	15.0	70	1	chevy c20
11	8	318.0	210	4382	13.5	70	1	dodge d200
9	8	304.0	193	4732	18.5	70	1	hi 1200d
27	4	97.0	88	2130	14.5	71	3	datsun pl510
28	4	140.0	90	2264	15.5	71	1	chevrolet vega 2300
25	4	113.0	95	2228	14.0	71	3	toyota corona
25	4	98.0	NA	2046	19.0	71	1	ford pinto
19	6	232.0	100	2634	13.0	71	1	amc gremlin
16	6	225.0	105	3439	15.5	71	1	plymouth satellite custom
17	6	250.0	100	3329	15.5	71	1	chevrolet chevelle malibu
19	6	250.0	88	3302	15.5	71	1	ford torino 500
18	6	232.0	100	3288	15.5	71	1	amc matador
14	8	350.0	165	4209	12.0	71	1	chevrolet impala
14	8	400.0	175	4464	11.5	71	1	pontiac catalina brougham
14	8	351.0	153	4154	13.5	71	1	ford galaxie 500
14	8	318.0	150	4096	13.0	71	1	plymouth fury iii
12	8	383.0	180	4955	11.5	71	1	dodge monaco (sw)
13	8	400.0	170	4746	12.0	71	1	ford country squire (sw)
13	8	400.0	175	5140	12.0	71	1	pontiac safari (sw)
18	6	258.0	110	2962	13.5	71	1	amc hornet sportabout (sw)
22	4	140.0	72	2408	19.0	71	1	chevrolet vega (sw)
19	6	250.0	100	3282	15.0	71	1	pontiac firebird
18	6	250.0	88	3139	14.5	71	1	ford mustang
23	4	122.0	86	2220	14.0	71	1	mercury capri 2000
28	4	116.0	90	2123	14.0	71	2	opel 1900
30	4	79.0	70	2074	19.5	71	2	peugeot 304
30	4	88.0	76	2065	14.5	71	2	fiat 124b
31	4	71.0	65	1773	19.0	71	3	toyota corolla 1200
35	4	72.0	69	1613	18.0	71	3	datsun 1200
27	4	97.0	60	1834	19.0	71	2	volkswagen model 111
26	4	91.0	70	1955	20.5	71	1	plymouth cricket
24	4	113.0	95	2278	15.5	72	3	toyota corona hardtop
25	4	97.5	80	2126	17.0	72	1	dodge colt hardtop
23	4	97.0	54	2254	23.5	72	2	volkswagen type 3
20	4	140.0	90	2408	19.5	72	1	chevrolet vega
21	4	122.0	86	2226	16.5	72	1	ford pinto runabout
13	8	350.0	165	4274	12.0	72	1	chevrolet impala
14	8	400.0	175	4385	12.0	72	1	pontiac catalina
15	8	318.0	150	4135	13.5	72	1	plymouth fury iii
14	8	351.0	153	4129	13.0	72	1	ford galaxie 500
17	8	304.0	150	3672	11.5	72	1	amc ambassador sst
11	8	429.0	208	4633	11.0	72	1	mercury marquis
13	8	350.0	155	4502	13.5	72	1	buick lesabre custom
12	8	350.0	160	4456	13.5	72	1	oldsmobile delta 88 royale
13	8	400.0	190	4422	12.5	72	1	chrysler newport royal
19	3	70.0	97	2330	13.5	72	3	mazda rx2 coupe
15	8	304.0	150	3892	12.5	72	1	amc matador (sw)
13	8	307.0	130	4098	14.0	72	1	chevrolet chevelle concours (sw)
13	8	302.0	140	4294	16.0	72	1	ford gran torino (sw)
14	8	318.0	150	4077	14.0	72	1	plymouth satellite custom (sw)
18	4	121.0	112	2933	14.5	72	2	volvo 145e (sw)
22	4	121.0	76	2511	18.0	72	2	volkswagen 411 (sw)
21	4	120.0	87	2979	19.5	72	2	peugeot 504 (sw)
26	4	96.0	69	2189	18.0	72	2	renault 12 (sw)
22	4	122.0	86	2395	16.0	72	1	ford pinto (sw)
28	4	97.0	92	2288	17.0	72	3	datsun 510 (sw)
23	4	120.0	97	2506	14.5	72	3	toyouta corona mark ii (sw)
28	4	98.0	80	2164	15.0	72	1	dodge colt (sw)
27	4	97.0	88	2100	16.5	72	3	toyota corolla 1600 (sw)
13	8	350.0	175	4100	13.0	73	1	buick century 350
14	8	304.0	150	3672	11.5	73	1	amc matador
13	8	350.0	145	3988	13.0	73	1	chevrolet malibu
14	8	302.0	137	4042	14.5	73	1	ford gran torino
15	8	318.0	150	3777	12.5	73	1	dodge coronet custom
12	8	429.0	198	4952	11.5	73	1	mercury marquis brougham
13	8	400.0	150	4464	12.0	73	1	chevrolet caprice classic
13	8	351.0	158	4363	13.0	73	1	ford ltd
14	8	318.0	150	4237	14.5	73	1	plymouth fury gran sedan
13	8	440.0	215	4735	11.0	73	1	chrysler new yorker brougham
12	8	455.0	225	4951	11.0	73	1	buick electra 225 custom
13	8	360.0	175	3821	11.0	73	1	amc ambassador brougham
18	6	225.0	105	3121	16.5	73	1	plymouth valiant
16	6	250.0	100	3278	18.0	73	1	chevrolet nova custom
18	6	232.0	100	2945	16.0	73	1	amc hornet

The model

auto_mpg<-na.omit(auto_mpg)

set.seed(1)
#lets split the data 60/40
library(caret)
trainIndex <- createDataPartition(auto_mpg$mpg, p = .6, list = FALSE, times = 1)

#grab the data
mpgTrain <- auto_mpg[ trainIndex,]
mpgTest  <- auto_mpg[-trainIndex,]

mpggbm<- train(
  form = mpg ~ cylinders+displacement+horsepower+weight+acceleration+factor(model_year)+factor(origin),
  data = mpgTrain,
  #here we add classProbs because we want probs
  trControl = trainControl(method = "cv", number = 10),
  method = "gbm",
  tuneLength = 10,
  verbose=FALSE)

knitr::kable(mpggbm$bestTune)

	n.trees	interaction.depth	shrinkage	n.minobsinnode
6	300	1	0.1	10

summary(mpggbm)

                                      var     rel.inf
displacement                 displacement 26.94621833
weight                             weight 23.17435112
horsepower                     horsepower 19.63108746
cylinders                       cylinders  7.06435813
factor(model_year)82 factor(model_year)82  5.46316709
factor(model_year)80 factor(model_year)80  4.57983291
acceleration                 acceleration  4.26451329
factor(model_year)81 factor(model_year)81  2.30193941
factor(model_year)73 factor(model_year)73  2.14169961
factor(model_year)79 factor(model_year)79  1.28966117
factor(model_year)72 factor(model_year)72  1.10759434
factor(origin)3           factor(origin)3  0.80155769
factor(origin)2           factor(origin)2  0.40108307
factor(model_year)78 factor(model_year)78  0.33288986
factor(model_year)74 factor(model_year)74  0.25780334
factor(model_year)76 factor(model_year)76  0.12238033
factor(model_year)71 factor(model_year)71  0.05678884
factor(model_year)77 factor(model_year)77  0.03278255
factor(model_year)75 factor(model_year)75  0.03029146

mpggbm_Pred<-predict(mpggbm,mpgTest)

knitr::kable(mpggbm_Pred)%>%
  kableExtra::kable_styling("striped")%>%
  kableExtra::scroll_box(width = "50%",height="300px")

x
17.22115
17.64381
14.93228
14.93228
15.09813
16.24625
16.39585
17.92315
19.77060
22.58967
24.99121
25.92369
14.10710
13.93441
17.91073
17.31234
15.08089
27.26531
24.03732
33.31186
31.93535
13.32985
12.79194
13.49570
14.52023
12.79194
24.74133
24.65089
28.73584
22.01284
25.18892
24.72307
12.27333
12.10748
12.12382
11.56958
18.26998
11.17757
11.71917
11.56958
24.56904
23.46576
19.89709
22.25626
12.27333
11.95788
27.59018
19.77070
14.45734
20.47107
17.63908
31.21116
25.62049
15.67514
16.55074
14.77456
14.24669
26.82490
34.65491
30.39243
17.44486
14.03970
18.16763
18.10036
23.89661
30.10882
30.04516
18.47023
15.27834
15.14153
15.03680
17.34817
20.87879
18.65120
31.71094
30.04482
31.30393
16.34144
16.39732
28.19232
21.07537
14.99837
20.53874
22.23192
29.89433
15.47921
14.93761
15.78728
19.24194
26.35599
24.39282
27.53039
32.07788
26.67017
27.60152
24.18956
24.61794
33.41975
28.74166
33.45902
32.15866
16.99719
21.86984
19.34578
18.41856
18.01985
27.25457
23.70732
26.08560
22.73485
21.74154
23.47636
21.02211
17.33256
17.78127
34.31357
28.16170
27.53156
17.92389
26.97398
32.70180
23.68881
27.85709
24.62995
32.11614
28.46064
31.50306
33.01996
38.21091
27.87489
30.16567
34.03440
35.77455
29.21421
33.76378
28.33990
29.30736
29.33523
35.57827
35.17441
37.26067
33.32268
35.60199
33.24434
25.95727
27.02805
26.75234
31.97432
28.87719
33.70910
38.26585
35.08431
26.15340
30.58842
32.14702

mpggbmtestpred<-cbind(mpggbm_Pred,mpgTest)

#root mean squared error
RMSE(mpggbmtestpred$mpggbm_Pred,mpggbmtestpred$mpg)

[1] 2.966667

#best measure ever...RSquared 
cor(mpggbmtestpred$mpggbm_Pred,mpggbmtestpred$mpg)^2

[1] 0.8578063

install.packages("pdp")

Gradient Tree Boosting: Partial Dependence

The feature importance shown earlier provides an indication of the relative importance of the different features, for either classification or regression. An alternate approach to understanding the relationship between features and their importance involves the construction of partial dependence plots. These plots show the dependence between the regression model and a specific set of features, selected from the set of all features used to build the regression model. To calculate this dependence, we marginalize over all of the other features, so that the effect of the target feature (or features) is measured. Generally, this set is restricted to a small number, such as one or two, since it becomes hard to visualize larger relationships.

pdp::partial(mpggbm, pred.var = "cylinders", plot = TRUE, rug = TRUE,
              plot.engine = "ggplot2")

gridExtra::grid.arrange(
  pdp::partial(mpggbm, pred.var = "cylinders", plot = TRUE, rug = TRUE,
              plot.engine = "ggplot2"),
  pdp::partial(mpggbm, pred.var = "displacement", plot = TRUE, rug = TRUE,
              plot.engine = "ggplot2"),
  ncol = 2 
)

  pdp::partial(mpggbm, pred.var = c("cylinders","displacement"), plot = TRUE, rug = TRUE,
              plot.engine = "ggplot2")

# Compute partial dependence data for lstat and rm
pd <- pdp::partial(mpggbm, pred.var = c("weight","displacement"))

# Default PDP
pdp::plotPartial(pd)

# Add contour lines and use a different color palette
rwb <- colorRampPalette(c("darkred", "white", "pink"))
pdp::plotPartial(pd, contour = TRUE, col.regions = rwb)

# 3-D surface
pdp::plotPartial(pd, levelplot = FALSE, zlab = "mpg", colorkey = TRUE, 
                    screen = list(z = -20, x = -60))

#install.packages("akima")
###### Interactive 3D partial dependence plot with coloring scale ######

# Interpolate the partial dependence values
dens <- akima::interp(x = pd$weight, y = pd$displacement, z = pd$yhat)

# 3D partial dependence plot with a coloring scale
p3 <- plotly::plot_ly(x = dens$x, 
          y = dens$y, 
          z = dens$z,
          colors = c("blue", "grey", "red"),
          type = "surface")
# Add axis labels for 3D plots

p3 <- p3%>% plotly::layout(scene = list(xaxis = list(title = 'Weight'),
                     yaxis = list(title = 'Displacement'),
                     zaxis = list(title = 'Predicted Mpg')))
# Show the plot
p3

Adaboost

We now turn our attention to the second boosting algorithm presented in this module, Adaboost, which is short for Adaptive Boosting. Adaboost was an early boosting algorithm that iteratively adjusts the weights on the training data to improve the performance of the ensemble algorithm. The weights are adjusted on each iteration to increase the predictive accuracy on the incorrectly predicted training data.

set.seed(1)
#lets split the data 60/40
iris2<-iris%>%
  filter(Species!="versicolor")
iris2$Species<-factor(iris2$Species)
  
library(caret)
trainIndex <- createDataPartition(iris2$Species, p = .6, list = FALSE, times = 1)

#grab the data
irisTrain <- iris2[ trainIndex,]
irisTest  <- iris2[-trainIndex,]

#added something here
Irisadaboost<- train(
  form = factor(Species) ~ .,
  data = irisTrain,
  #here we add classProbs because we want probs
  trControl = trainControl(method = "cv", number = 10,
                           classProbs =  TRUE),
  method = "adaboost",
  tuneLength = 5,
  #add this please
  verbose=FALSE)

#Irisadaboost
knitr::kable(Irisadaboost$bestTune)

nIter	method
50	Adaboost.M1

plot(Irisadaboost)

Irisadaboost_Pred<-predict(Irisadaboost,irisTest,type="prob")

knitr::kable(Irisadaboost_Pred)%>%
  kableExtra::kable_styling("striped")%>%
  kableExtra::scroll_box(width = "50%",height="300px")

setosa	virginica
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1

irisadaboosttestpred<-cbind(Irisadaboost_Pred,irisTest)

irisadaboosttestpred<-irisadaboosttestpred%>%
  mutate(prediction=if_else(setosa>virginica,"setosa",
                            if_else(virginica>setosa,"virginica", "PROBLEM")))

table(irisadaboosttestpred$prediction)


   setosa virginica 
       20        20

adaboostConfusion<-confusionMatrix(factor(irisadaboosttestpred$prediction),factor(irisadaboosttestpred$Species))

adaboostConfusion

Confusion Matrix and Statistics

           Reference
Prediction  setosa virginica
  setosa        20         0
  virginica      0        20
                                     
               Accuracy : 1          
                 95% CI : (0.9119, 1)
    No Information Rate : 0.5        
    P-Value [Acc > NIR] : 9.095e-13  
                                     
                  Kappa : 1          
                                     
 Mcnemar's Test P-Value : NA         
                                     
            Sensitivity : 1.0        
            Specificity : 1.0        
         Pos Pred Value : 1.0        
         Neg Pred Value : 1.0        
             Prevalence : 0.5        
         Detection Rate : 0.5        
   Detection Prevalence : 0.5        
      Balanced Accuracy : 1.0        
                                     
       'Positive' Class : setosa

setosa	virginica
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1

setosa	virginica
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
1	0
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1
0	1

Gradient Boosting