Abstract: Since the mid-90s, utilization of the web has augmented in different structures. Individuals are conveying utilizing different appearances. With this humungous development of web traffic, distinctive online web- based platforms, for example, Facebook, Twitter, LinkedIn, and so forth, are likewise getting popular. Billions of clients are imparting their insight on various perspectives on websites, for example, Facebook, Tumbler, Twitter, glint, LinkedIn, and so forth. Twitter is the most well-known small-scale blogging and person to person communication administration, which gives an office to clients to sharing, conveying, and translating 140 words' posts known as a tweet. Twitter has 320 Million month to month dynamic clients. It is open through site interface, SMS, or portable gadgets. Characteristic language handling is additionally assuming a major job and can be utilized by the suppositions communicated. In this project, the authors are trying to analyze the users’ sentiment on tweets collected from different sources on the IMDB movie reviews. The project successfully predicts if the sentiment of the review will be of a negative or positive polarity. In this paper, the authors came across different methods of implementations and identified various techniques through which features of the reviews can be defined considering it either as single entities or a whole document. The authors of this project implemented a technique called distributed representation of sentences and documents, which ultimately gave them the highest accuracy among all the models they implemented.

Keywords: Machine Learning, Sentiment Analysis, IMDB, Random Forest, ANN, SVM

PDF | DOI: 10.17148/IJARCCE.2020.9522

Open chat
Chat with IJARCCE