Book Image

Mastering .NET Machine Learning

By : Jamie Dixon, Damian R Mingle
Book Image

Mastering .NET Machine Learning

By: Jamie Dixon, Damian R Mingle

Overview of this book

.Net is one of the widely used platforms for developing applications. With the meteoric rise of Machine learning, developers are now keen on finding out how can they make their .Net applications smarter. Also, .NET developers are interested into moving into the world of devices and how to apply machine learning techniques to, well, machines. This book is packed with real-world examples to easily use machine learning techniques in your business applications. You will begin with introduction to F# and prepare yourselves for machine learning using .NET framework. You will be writing a simple linear regression model using an example which predicts sales of a product. Forming a base with the regression model, you will start using machine learning libraries available in .NET framework such as Math.NET, Numl.NET and Accord.NET with the help of a sample application. You will then move on to writing multiple linear regressions and logistic regressions. You will learn what is open data and the awesomeness of type providers. Next, you are going to address some of the issues that we have been glossing over so far and take a deep dive into obtaining, cleaning, and organizing our data. You will compare the utility of building a KNN and Naive Bayes model to achieve best possible results. Implementation of Kmeans and PCA using Accord.NET and Numl.NET libraries is covered with the help of an example application. We will then look at many of issues confronting creating real-world machine learning models like overfitting and how to combat them using confusion matrixes, scaling, normalization, and feature selection. You will now enter into the world of Neural Networks and move your line of business application to a hybrid scientific application. After you have covered all the above machine learning models, you will see how to deal with very large datasets using MBrace and how to deploy machine learning models to Internet of Thing (IoT) devices so that the machine can learn and adapt on the fly.
Table of Contents (18 chapters)
Mastering .NET Machine Learning
Credits
About the Author
Acknowledgments
About the Reviewers
www.PacktPub.com
Preface
Index

Overfitting and cross validation


If you remember from Chapters 2, 3, and 4, one of the problems with our methodology when building models was that we were guilty of overfitting. Overfitting, the bane of predictive analytics, is what happens when we build a model that does a great job with past data but then falls apart when new data is introduced. This phenomenon is not just for data science; it happens a lot in our society: Professional athletes get lucrative contracts and then fail to live up to their prior performances; fund managers get hefty salary bumps because of last year's performance, and the list goes on.

Cross validation – train versus test

Unlike the Yankees, who never seem to learn, our profession has learned from its mistakes and has a great, if imperfect, tool to combat overfitting. We use the methodology of train/test/eval to build several models and then select the best one not based on how well it did against an existing dataset, but how it does against data it has never...