Search icon CANCEL
Subscription
0
Cart icon
Cart
Close icon
You have no products in your basket yet
Save more on your purchases!
Savings automatically calculated. No voucher code required
Arrow left icon
All Products
Best Sellers
New Releases
Books
Videos
Audiobooks
Learning Hub
Newsletters
Free Learning
Arrow right icon
Arrow up icon
GO TO TOP
Mastering Machine Learning with R, Second Edition - Second Edition

You're reading from  Mastering Machine Learning with R, Second Edition - Second Edition

Product type Book
Published in Apr 2017
Publisher Packt
ISBN-13 9781787287471
Pages 420 pages
Edition 2nd Edition
Languages
Toc

Table of Contents (23) Chapters close

Title Page
Credits
About the Author
About the Reviewers
Packt Upsell
Customer Feedback
Preface
1. A Process for Success 2. Linear Regression - The Blocking and Tackling of Machine Learning 3. Logistic Regression and Discriminant Analysis 4. Advanced Feature Selection in Linear Models 5. More Classification Techniques - K-Nearest Neighbors and Support Vector Machines 6. Classification and Regression Trees 7. Neural Networks and Deep Learning 8. Cluster Analysis 9. Principal Components Analysis 10. Market Basket Analysis, Recommendation Engines, and Sequential Analysis 11. Creating Ensembles and Multiclass Classification 12. Time Series and Causality 13. Text Mining 14. R on the Cloud 15. R Fundamentals 16. Sources

Modeling evaluation and selection


As we've done in prior chapters, the first recommended task when utilizing caret functions is to build the object that specifies how model training is going to happen. This is done with the trainControl() function. We are going to create a five-fold cross-validation and save the final predictions (the probabilities). It is recommended that you also index the resamples so that each base model trains on the same folds. Also, notice in the function that I specified upsampling. Why? Well, notice that the ratio of "Yes" versus "No" is 2 to 1:

    > table(train$type)

     No Yes 
    267 133

This ratio is not necessarily imbalanced, but I want to demonstrate something here. In many data sets, the outcome of interest is a rare event. As such, you can end up with a classifier that is highly accurate but does a horrible job at predicting the outcome of interest, which is to say it doesn't predict any true positives. To balance the response, you can upsample the...

lock icon The rest of the chapter is locked
Register for a free Packt account to unlock a world of extra content!
A free Packt account unlocks extra newsletters, articles, discounted offers, and much more. Start advancing your knowledge today.
Unlock this book and the full library FREE for 7 days
Get unlimited access to 7000+ expert-authored eBooks and videos courses covering every tech area you can think of
Renews at $15.99/month. Cancel anytime