reqspec.md

In AdvancedR-2021/hmm: Hidden Markov Models

Tentative contents of the package

Functions

• Forward algorithm for HMM's
• Backward algorithm for HMM's
• EM-algorithm for non-stationary HMM's, where users can either choose from standard distributions (Poisson, Normal etc.) or provide custom density and estimation functions
• Direct likelihood optimization for HMM's
• Functions for prediction: Decoding, outlier detection, forecasting etc.

Classes

Mainly one hmm class similar to that of lm, including many of the same methods such as plot, summary, predict, simulate etc.

Data

Different simulated data-sets using different marginal distributions.

Perhaps also fitting real-world data, if any such is freely available.

AdvancedR-2021/hmm documentation built on Dec. 17, 2021, 7:41 a.m.