Mastering .NET Machine Learning电子书

Mastering .NET Machine Learning

Table of Contents

Mastering .NET Machine Learning

Credits

About the Author

Acknowledgments

About the Reviewers

www.PacktPub.com

eBooks, discount offers, and more

Why subscribe?

Preface

What this book covers

What you need for this book

Who this book is for

Conventions

Reader feedback

Customer support

Downloading the example code

Errata

Piracy

Questions

1. Welcome to Machine Learning Using the .NET Framework

What is machine learning?

Why .NET?

What version of the .NET Framework are we using?

Why write your own?

Why open data?

Why F#?

Getting ready for machine learning

Setting up Visual Studio

Learning F#

Third-party libraries

Math.NET

Accord.NET

Numl

Summary

2. AdventureWorks Regression

Simple linear regression

Setting up the environment

Preparing the test data

Standard deviation

Pearson's correlation

Linear regression

Math.NET

Regression try 1

Regression try 2

Accord.NET

Regression

Regression evaluation using RMSE

Regression and the real world

Regression against actual data

AdventureWorks app

Setting up the environment

Updating the existing web project

Implementing the regression

Summary

3. More AdventureWorks Regression

Introduction to multiple linear regression

Intro example

Keep adding x variables?

AdventureWorks data

Adding multiple regression to our production application

Considerations when using multiple x variables

Adding a third x variable to our model

Logistic regression

Intro to logistic regression

Adding another x variable

Applying a logistic regression to AdventureWorks data

Categorical data

Attachment point

Analyzing results of the logistic regression

Adding logistic regression to the application

Summary

4. Traffic Stops – Barking Up the Wrong Tree?

The scientific process

Open data

Hack-4-Good

FsLab and type providers

Data exploration

Visualization

Decision trees

Accord

numl

Summary

5. Time Out – Obtaining Data

Overview

SQL Server providers

Non-type provider

SqlProvider

Deedle

MicrosoftSqlProvider

SQL Server type provider wrap up

Non SQL type providers

Combining data

Parallelism

JSON type provider – authentication

Summary

6. AdventureWorks Redux – k-NN and Naïve Bayes Classifiers

k-Nearest Neighbors (k-NN)

k-NN example

Naïve Bayes

Naïve Bayes in action

One thing to keep in mind while using Naïve Bayes

AdventureWorks

Getting the data ready

k-NN and AdventureWorks data

Naïve Bayes and AdventureWorks data

Making use of our discoveries

Getting the data ready

Expanding features

Summary

7. Traffic Stops and Crash Locations – When Two Datasets Are Better Than One

Unsupervised learning

k-means

Principle Component Analysis (PCA)

Traffic stop and crash exploration

Preparing the script and the data

Geolocation analysis

PCA

Analysis summary

The Code-4-Good application

Machine learning assembly

The UI

Adding distance calculations

Augmenting with human observations

Summary

8. Feature Selection and Optimization

Cleaning data

Selecting data

Collinearity

Feature selection

Normalization

Scaling

Overfitting and cross validation

Cross validation – train versus test

Cross validation – the random and mean test

Cross validation – the confusion matrix and AUC

Cross validation – unrelated variables

Summary

9. AdventureWorks Production – Neural Networks

Neural networks

Background

Neural network demo

Neural network – try #1

Neural network – try #2

Building the application

Setting up the models

Building the UX

Summary

10. Big Data and IoT

AdventureWorks and the Internet of Bikes

Data considerations

MapReduce

MBrace

Distributed logistic regression

The IoT

PCL linear regression

Service layer

Universal Windows app and Raspberry Pi 2

Next steps

Summary

Index