Skip to content
/ hmm Public

Hidden Markov Model

License

Apache-2.0, MIT licenses found

Licenses found

Apache-2.0
LICENSE-APACHE
MIT
LICENSE-MIT
Notifications You must be signed in to change notification settings

klangner/hmm

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

20 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Hidden Markov Model

Crates.io Crates.io docs.rs

hmm is an early-stage open-source project. It means that API can change at any time. If you think that this library can help you, then let me know. We can discuss future direction and try to stabilize the API.

Features

  • Viterbi MAP estimation

Example

Lets say that we have 2 coins:

  • Fair which generates H (Head) and T (Tails) with probability of 1/2
  • Biased - with probabilities H: 1/4, T: 3/4

We also know that after each toss we can switch coin with the probability of

  • Use the same coin: 3/4
  • Switch coin: 1/4

First time we select coin with probability of 1/2

Using this library we can answer the following question:

Given the observations 'H H T T T' which coins were used during each toss?

Lest build HMM model and check the answer:

extern crate hmm;

use hmm::models::{HiddenMarkov};


fn main() {
    // 50%:50% chance to start at either coin
    let initials = vec![0.5, 0.5];

    // state transitions between coins:
    //   3/4 to stay on same coin (on matrix diagonals)
    //   1/4 chance to switch
    let st = vec![ vec![0.75, 0.25],
                   vec![0.25, 0.75]];

    // observation probabilities for each coin:
    //    first coin is fair, second one is biased
    let obs = vec![ vec![0.5, 0.5],
                    vec![0.25, 0.75]];

    // generate the HMM model
    let hmm = HiddenMarkov::new(initials, st, obs).unwrap();
    let coins = hmm.map_estimate(vec![0, 0, 1, 1, 1]);
    println!("Coins used: {:?}", coins);
}

For more check Examples.

License

Licensed under either of

at your option.

Contributions

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.

Releases

No releases published

Packages

No packages published

Languages