vinsis / word-classification

Using NLTK's Bayesian Classification to classify any roman string as English, Spanish or Japanese

word-classification

Accuracy: ~80%

Overview

A piece of code that tells classifies any given word sounds as Japanese, English or Spanish based on its structure.

For example, it classifies:

• "mirodo" as Japanese
• "mirado" as Spanish
• "mirror" as English

It is able to detect subtle differences between words. For example, it classifies "Texas" as English but "Tejas" as Spanish. Texas is actually anglicized form of the Spanish word Tejas. It also classifies Colorado as Spanish.

How to use it

In any Python Interpreter, go to the location where classifier.py is located and type the commands like so:

import classifier
classifier.trainData()
classifier.classifyWord('hello')