https://github.com/prml/prmlt

Matlab code of machine learning algorithms in book PRML
https://github.com/prml/prmlt

algorithms machine-learning machine-learning-algorithms matlab prml

Last synced: 3 days ago
JSON representation

Matlab code of machine learning algorithms in book PRML

• Host: GitHub
• URL: https://github.com/prml/prmlt
• Owner: PRML
• License: mit
• Created: 2012-09-10T08:19:18.000Z (over 12 years ago)
• Default Branch: master
• Last Pushed: 2020-03-04T13:50:30.000Z (almost 5 years ago)
• Last Synced: 2024-10-29T15:29:11.519Z (about 2 months ago)
• Topics: algorithms, machine-learning, machine-learning-algorithms, matlab, prml
• Language: MATLAB
• Homepage: http://prml.github.io/
• Size: 494 KB
• Stars: 6,025
• Watchers: 405
• Forks: 2,158
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
Introduction

-------

This Matlab package implements machine learning algorithms described in the great textbook:

Pattern Recognition and Machine Learning by C. Bishop ([PRML](http://research.microsoft.com/en-us/um/people/cmbishop/prml/)).

It is written purely in Matlab language. It is self-contained. There is no external dependency.

Note: this package requires Matlab **R2016b** or latter, since it utilizes a new Matlab syntax called [Implicit expansion](https://cn.mathworks.com/help/matlab/release-notes.html?rntext=implicit+expansion&startrelease=R2016b&endrelease=R2016b&groupby=release&sortby=descending) (a.k.a. broadcasting). It also requires Statistics Toolbox (for some simple random number generator) and Image Processing Toolbox (for reading image data).

Design Goal

-------

* Succinct: The code is extremely compact. Minimizing code length is a major goal. As a result, the core of the algorithms can be easily spotted.

* Efficient: Many tricks for speeding up Matlab code are applied (e.g. vectorization, matrix factorization, etc.). Usually, functions in this package are orders faster than Matlab builtin ones (e.g. kmeans).

* Robust: Many tricks for numerical stability are applied, such as computing probability in logrithm domain, square root matrix update to enforce matrix symmetry\PD, etc.

* Readable: The code is heavily commented. Corresponding formulas in PRML are annoted. Symbols are in sync with the book.

* Practical: The package is not only readable, but also meant to be easily used and modified to facilitate ML research. Many functions in this package are already widely used (see [Matlab file exchange](http://www.mathworks.com/matlabcentral/fileexchange/?term=authorid%3A49739)).

Installation

-------

1. Download the package to a local folder (e.g. ~/PRMLT/) by running: 

```console

git clone https://github.com/PRML/PRMLT.git

```

2. Run Matlab and navigate to the folder (~/PRMLT/), then run the init.m script.

3. Run some demos in ~/PRMLT/demo folder. Enjoy!

FeedBack

-------

If you find any bug or have any suggestion, please do file issues. I am graceful for any feedback and will do my best to improve this package.

License

-------

Released under MIT license

Contact

-------

sth4nth at gmail dot com