Introduction to Statistical Machine Learning with MicrosoftML
HTML Other
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.

README.md

Statistical Machine Learning with MicrosoftML

Description

This workshop covers the fundamentals of statistical machine learning with the MicrosoftML package. MicrosoftML provides a R interface to a set of scalable and distributed learning algorithms and data transformers. The package was initially developed by the research team at Microsoft (MSR), and powers the majority of machine learning applications within Microsoft and within the Azure cloud machine learning ecosystem. In this workshop, you will learn how you can use MicrosoftML’s state-of-the-art machine learning capabilities within R to train high-accuracy machine learning algorithms at blazingly fast speed. We’ll learn the core syntax of the package, how to use it in tandem with your favorite packages for tidy data processing and visualization, and finally how to deploy your trained algorithms in production environments (including Spark clusters).

MRS

MicrosoftML is a package that works in tandem with the RevoScaleR package and Microsoft R Server. In order to use the MicrosoftML and RevoScaleR libraries, you need an installation of Microsoft R Server or Microsoft R Client. You can download Microsoft R Server through MSDN here:

  1. R Server for Linux
  2. R Server for Hadoop
  3. R Server for Windows
  4. R Server for SQL Server (In-Database)

You can download Microsoft R Client through the following sources:

  1. R Client for Windows
  2. R Client for Linux
  3. R Client Docker Image

MicrosoftML References

Intended Audience

This course will be useful to anyone looking to use the advanced analytics capabilities provided by Microsoft ML Server. A minimal background in R and statistics are assumed, which you can refresh with the resources below.

Syllabus

  1. Exploratory Data Analysis and Feature Engineering
  2. Training Regression Models with ML Server
  3. Classification Models for Computer Vision
  4. Convolutional Neural Networks for Computer Vision
  5. Natural Language Processing and Text Classification
  6. Transfer Learning with Pre-Trained Deep Neural Network Architectures – The Shallow End of Deep Learning

Prerequisities

This course assumes some R background.

Introductory Courses

Useful Resources