Practices for Advanced Java Course

Practice 1 (WordSeg)

A simple Chinese segmentation system based on Hidden Markov Model.

Abstract:

1. train model from data/train
2. test model by data/test
3. test model by commandline input
4. save model to files(A.matrix, B.matrix, Cnt.matrix)
5. load model from files

Dataset:

Second International Chinese Word Segmentation Bakeoff Data

References:

Hidden Markov Model

Viterbi Algorithm