kimrojas / prml_python

Implement prml contents with python

PRML

These are codes implementing some algorithms introduced in "Pattern Recognition and Machine Learning" (Author: C.M.Bishop). Python language used for these implementation.

Required packages

• python 3
• numpy
• pandas
• scipy
• matplotlib

Installation

1. Download the file to a local folder (e.g. ~/prml_python/) by executing:

git clone https://github.com/oilneck/prml_python.git

2. Run Python and change your directory (~/prml_python/), then run the init.py script.

3. Run some demonstration files in Chap1~Chap11 folder.

Execution example

chap.1 : bayesian fitting	chap.4 : logistic regression	chap.5 : neural network

Jupyter Notebook files

The contents of Pattern Recognition and Machine Learning

• chap.1 : Introduction
• chap.2 : Probability distributions
• chap.3 : Linear Models for Regression
• chap.4 : Linear Models for Classification
• chap.5 : Neural Networks
• chap.6 : Kernel Methods
• chap.7 : Sparse Kernel Machines
• chap.8 : Graphical Models
• chap.9 : Mixture Models and EM
• chap.10 : Approximate Inference
• chap.11 : Sampling Methods

Deep learning and Convolutional neural network for image recognition

• Image recognition 【 Required libraries : keras, TensorFlow, OpenCV 】
• Deep learning test 【 Required libraries : numpy, sklearn (←to fetch data) 】
• Sequential models 【 Required libraries : numpy 】

NOTICE

All sources in [~/prml_python/prml] are the module file. If you want to change certain parameters (ex. iteration number, activation function in each layer for Neural Network), check the files in that directory.

External links

Wiki	Wiki for prml algorithm
Text	Regularization_of_NN.pdf
Slide	CNN.pdf