Machine learning techniques have become increasingly popular and relevant to solve text and sentiment related problems in recent years. It has boosted performance on several tasks and significantly reduced the necessity for human efforts. For this project, we will focus on text classification, especially sentiment analysis, on several datasets. Since there are researches about sentiment analysis on Amazon Review dataset, we will first use the same methods on the same dataset, and then replicate the methods on datasets other than the Amazon Reviews dataset. By completing the project, we are trying to realize the following goals:
- Build four text sentiment classifiers on Amazon Review dataset (BoW, Word2Vec, GloVe, fastText);
- Re-apply the above four classifiers on Drug Reveiw dataset;
- Compare different classifiers for datasets
rongl2: Data preprocessing, visualization, apply word2vec on Amazon review and Drug review dataset. (33.3%)
xdai12: Data preprocessing, visualization, apply GloVe on Amazon review and Drug review dataset, and building rmarkdown file. (33.3%)
zixingd2: BoW and FastText on Amazon review and Drug review dataset. (33.3%)