mahout-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From conflue...@apache.org
Subject [CONF] Apache Mahout > Parallel Viterbi
Date Sun, 11 Sep 2011 19:14:01 GMT
Space: Apache Mahout (https://cwiki.apache.org/confluence/display/MAHOUT)
Page: Parallel Viterbi (https://cwiki.apache.org/confluence/display/MAHOUT/Parallel+Viterbi)

Added by Sergey Bartunov:
---------------------------------------------------------------------
Viterbi algorithm is known as inference algorithm (synonyms: segmentation, decoding etc) for
Hidden Markov Model \[1\] which finds the most likely sequence of hidden states by given sequence
of observed states.

Apache Mahout has both [sequential|Hidden Markov Model] and parallel (that's what you're reading
about) implementations of the algorithm.

h3. Parallelization strategy

is quite straightforward. The algorithm is

h3. Running the code

Run "bin/apache pviterbi" and see what it wants from you. That is: 

* serialized HmmModel (i.e. by LossyHmmModelSerializer class)
* input data (observed sequences) in the format described above



*References*

# [Wikipedia article|http://en.wikipedia.org/wiki/Viterbi_algorithm]

Change your notification preferences: https://cwiki.apache.org/confluence/users/viewnotifications.action

Mime
View raw message