Efficient Comment Classification through NLP and Fuzzy Classification

A significant increase has been noticed in the number of people that are utilizing the internet paradigm for various purposes such as accessing various portals such as Social media and E commerce websites. Due to their immense popularity, these platforms have seen remarkable growth and an increasing user base that is constantly interacting on the platform through the use of comments. These comments are mostly the users assisting each other on the platform in making the right decision. These comments can range from helpful to sarcastic, which can be highly difficult for a Natural Language Processing platform to determine. Supervised machine learning approach requires labels such as star ratings in reviews to understand the reviews and classify. These labels need to be reliable, whereas as they are entered by users, they could be misleading. Therefore, in this paper, an unsupervised approach towards the automatic classification of comments has been outlined in much detail. The proposed methodology utilizes an innovative combination of Term Frequency – Inverse Document Frequency TF IDF in addition to the NLP paradigm along with the addition of the Entropy Estimation through Shannon Information gain. This procedure can effectively disintegrate the sentence into its basic form which can then ultimately be classified using the Fuzzy Classification technique. Shubham Derhgawen | Himaja Gogineni | Subhasish Chatterjee | Rajesh Tak "Efficient Comment Classification through NLP and Fuzzy Classification" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-4 | Issue-3 , April 2020, URL: https://www.ijtsrd.com/papers/ijtsrd30758.pdf