Clustering Analogous Words in Myanmar Language using Word Embedding Model

Abstract

Word embedding represents the words in terms of vectors. It is influenced on different NLP research areas such as document classification, author identification, sentiment analysis, etc. One of the most popular embedding models is Word2Vec model. It provides efficient representations of words by using Continuous Bag of Words model (CBOW) and Skip Gram model. In English language, word embedding model can be applied for data preprocessing well but there is a very little amount of work done in Myanmar language. Text preprocessing is important part to build embedding model and it is a significantly effect on final results. This paper tries to extract the analogous words between Myanmar news articles focus on the bag of words (CBOW) model using different features vector sizes. By analyzing word embedding model are obtained the better results with a high dimensional vectors than a low dimensional vectors to cluster the words based on its relatedness.