Myan-word-breaker

Word segmentation tool for Myanmar sentence.

Demo

Try this library at flask-py-word-breaker

Introduction

This is the word segmentation tool for Myanmar text. Word segmentation is the process of determining word boundaries in a piece of text.

Word Segmentation is the most basic but very important step for Natural Language Processing of Myanmar Text. It is also non-trivial task since Myanmar text is a string of characters without explicit word boundary delimiters.

In this library, the method from the research paper of Ko Tun Thura Thet (2008) Word segmentation for the Myanmar Language is used.

The library supports two ways to find the possible combinations

All possible combinations loop
Sub-word possibility recursive method.

The first one might be a little bit better to provide more precise result while the later one has huge improvement in runtime performance.

To validate the evaluate these possible combinations, the library uses Bigram Collocation Strength Statistical Approach according to described in the research paper.

Usage

# coding=utf8
from word_breaker.word_segment_v5 import WordSegment

wordSegmenter = WordSegment()
# Segment using sub_word_possibility segmentation method on Unicode String
print(wordSegmenter.normalize_break('သဘာဝဟာသဘာဝပါ', 'unicode', wordSegmenter.SegmentationMethod.sub_word_possibility))

# Segment using sub_word_possibility segmentation method on Zawgyi String
print(wordSegmenter.normalize_break('သဘာဝဟာသဘာဝပါ', 'zawgyi', wordSegmenter.SegmentationMethod.sub_word_possibility))

# Segment using all_possible_combination segmentation method on Unicode String
print(wordSegmenter.normalize_break('သဘာဝဟာသဘာဝပါ', 'unicode', wordSegmenter.SegmentationMethod.all_possible_combination))

# Segment using all_possible_combination segmentation method on Zawgyi String
print(wordSegmenter.normalize_break('သဘာဝဟာသဘာဝပါ', 'zawgyi', wordSegmenter.SegmentationMethod.all_possible_combination))

myanmarlinks / myan-word-breaker Goto Github PK