Stochastic Modeling in R

Introduction

stochmod is an R library that implements learning and inference algorithms for some of the most popular probabilistic models and features:

• Linear and quadratic discriminant analysis algorithms
• Modeling with Gaussian Mixture Models, including:
  • learning via the Expectation Maximization (EM) algorithm
• Modeling with Hidden Markov Models, including:
  • learning via the EM algorithm
  • inference of the most likely state sequence using the Viterbi algorithm

The library features redundant implementation in two languages, R and C. The user controls which implementation is being executed through a simple setting of a flag. The R implementation has been optimized for readability and can be used for step-by-step educational purposes. The C implementation has been optimized for speed and is executed by default. stochmod has been designed using the functional programming paradigm. As a result, it is very easy and intuitive to use.

The package documentation includes my e-mail address. Please use it to report bugs.

Documentation

Requirements

stochmod library requires R version 2.4.0 or later.

Also, the following R packages are used by stochmod and must be installed a priori:

• mvtnorm

Installation

1. Download the package to your local drive
2. Install the package using the following command:
   R CMD INSTALL stochmod_1.2.tar.gz

Download

Local:

• Version 1.2 (changes)
• Version 1.1 (changes)
• Version 1.0