EarlGlynn / kc-r-users-caret-2017

Talk given on Sept. 9, 2017 to KC R Users Group: Using R's Caret Package for Machine Learning

Using R's Caret Package for Machine Learning

Talk given on Sept. 9, 2017 to KC R Users Group.

After giving an overview of the Caret Package, machine learning results using the caret package are used to compare performance of about a dozen models. The Forensic Glass dataset from the MASS package is used in most of the examples.

Powerpoint

Caret-Machine-Learning.pptx
Caret-Machine-Learning.pdf

Machine Learning Examples using Caret

R Markdown and corresponding HTML files: Forensic-Glass-caret-FILE.Rmd and Forensic-Glass-caret-FILE.html

FILE

C50
glmnet-SMOTE
J48
knn
lda
lda-YeoJohnson
nb
nb-ica
nnet
rf
rf-SMOTE
rpart
svmRadial

See the Results Summary slide for a short description of each example.

Related Examples

caret-overview.Rmd and .html:  Caret Model Summary and examples of getting info about particular models.

Forensic-Glass-Exploratory.Rmd and .html:  Examples of caret's visualizations.

Forensic-Glass-Heatmap-Clustering.Rmd and .html:  Views of Forensic Glass data as a heatmap

Examples moved

These files are now part of Survey of Machine Learning Feature Selection Methods talk:

Forensic-Glass-Correlation.Rmd and .html:  Correlation matrix of forensic glass predictors.

Forensic-Glass-caret-glmnet.Rmd and .html:  Elastic-Net example.

Forensic-Glass-Boruta.Rmd and .html:  Boruta 'All Relevant' Variables

Forensic-Glass-SVD.Rmd and .html:  Singular Value Decomposition
Forensic-Glass-PCA.Rmd and .html:  Principal Component Analysis (creates animated GIF with 3D view of first 3 PCs).

Toy Shiny App

The ShinyCaret folder contains a Shiny app originally created in 2014 for the Johns Hopkins Coursera class, Developing Data Products.

The app is not very practical given the run time for some of the algorithms, but the example shows how caret can be used in a Shiny app.