By Arnav Bhutani

The goal of bhutania-project-viterbi is to implement the viterbi algorithm in R using an example path through a Hidden Markov Model. This approach was done using a combination of the Baum-Welch Algorithm and the Viterbi Algorithm.

Project Report

Project Walkthrough

main file

Required Packages:

• testthat
• devtools
• HMM (for testing/timing)
• BenchMark
• Tidyverse
• Rcpp
• Here
• Readr

Important Code Files:

• main.R

Primary Functions:

• Forward: This function will be called by the EM algorithm, BaumWelch to find the probability of the next hidden state from a given visible state.
  • Usage: forward({A path of visible states}, {switching probabilties}, {visible (emission) probabilities}, {probability of observing first hidden state})
• Backward: This function will be called by the EM algorithm, BaumWelch to find the probability of the previous hidden state from a previous visible state.
  • Usage: (same as forward) backward({a path of visible states}, {switching probabilties}, {visible probabilities}, {probability of observing first hidden state})
• BaumWelch: This function uses expectation-maximization to find the largest probability for the switching and visible(emission) probabities given a visible path and some initial values.
  • Usage: BaumWelch({a set of visible states}, {initial switching probability}, {initial visible probability}, {initial distribution of hidden states}, {number of iterations})
• viterbi.R: This function uses the probabilities derived from the BaumWelch algorithm to find the most likely hidden state for each visible state in a path trace.
  • Usage: viterbi({set of visible states}, {switching probability}, {visible probability}, {initial distribtion of hidden states})