2026-09-20T21:02:57Z https://meral.edu.mm/oai

oai:meral.edu.mm:recid/6332 2022-03-24T23:17:19Z 1582963342780:1605779935331 user-uit

Sentiment Aware Word Embedding Approach for Sentiment Analysis Win Lei Kay Khine Nyein Thwet Thwet Aung Nowadays, many business owners want to know the feedback of their products. If they get the feedback from customers, they can promote the quality of their products. So, Sentiment analysis has become a popular research problem to tackle in NLP field. It is the process of identifying whether the opinion or reviews expressed in a piece of work is positive, negative or neutral. We can apply sentiment analysis in brand monitoring, customer service, market research and analysis. Word embedding step is a problem in sentiment analysis of neural network models. Most existing algorithms for continuous word representation typically only model the syntactic context of words but ignore the sentiment of text. It is a problematic for sentiment analysis as they usually map words with similar syntactic context but ignore opposite sentiment polarity, such as good and bad, like and dislike. We solve this issue by proposing a method, sentiment-aware word embedding (SAWE). SAWE encodes sentiment information in the continuous representation of words by using (1) prediction the model and (2) ranking model. Finally, we evaluate our proposed method on IMDB movie review and twitter datasets, after that we prove our method outperform than other word embedding methods like word2vec and GloVe. 2018-11-02 http://hdl.handle.net/20.500.12678/0000006332 https://meral.edu.mm/records/6332