期刊名称：International Journal of Advances in Engineering and Management
电子版ISSN：2395-5252
出版年度：2022
卷号：4
期号：5
页码：726-731
DOI：10.35629/5252-0405428433
语种：English
出版社：IJAEM JOURNAL
摘要：Sentiment analysis is one of the fastest growing research place, which enables customers to make better-informed purchase choices through right understanding and analysis of collective sentiments from the net and social media. It also presents organizations the capability to measure the impact in their social advertising strategies by figuring out the public emotions in the direction of the product or the occasions associated to them. Maximum of the research done thus far have targeted on obtaining sentiment features by analyzing syntactic and lexical features that had been explicitly expressed via sentiment words, emoticons and other special symbols. An approach is proposed to perform the sentiment analysis of product reviews using Deep Learning. Unlike traditional machine studying methods, Deep Learning models do not depend upon feature extractors as these features are learned directly throughout the training procedure. The main idea on this work is to use word2vec to research phrase embedding and recurrent neural networks to train and classify the sentiment lessons of the product evaluations. This blended word2vec- Long Short Term Memory version may be used to predict the sentiment of new product reviews. The proposed work ambitions to measure the accuracy of the sentiment analysis class version the usage of deep getting to know and neural networks.
关键词：NLP(Natural Language Processing);Sentiment Analysis;LSTM(Long Short Term Memory);Word vector model;Deep Learning;Word Embedding
