Cognitive Behavior and Human Computer Interaction Based on Machine Learning Algorithms

14Comparative Sentiment Analysis Through Traditional and Machine Learning-Based Approach

Sandeep Singh^* and Harjot Kaur

Department of Computer Science, Guru Nanak Dev University, Gurdaspur Campus, Punjab, India

*Corresponding author: [e-mail]

Abstract

People’s opinions and sentiments toward products, organizations, and their services can be evaluated through a classification of texts, speeches, and sign-language technologies. Natural Language Processing (NLP) used to perform such subjective tasks is Sentiment Analysis (SA). [1] coined the term SA for the first time finding positive or negative polarities of texts written by people. Nowadays, everybody reads user-review texts for evaluating the sentiments of users before buying any new product. However, it is cumbersome tasks to interpret the exact type of SA, as there are various kinds of SA available, out of which four are essential. These are fine-grained SA, emotion-detection, aspect-based, and intent-analysis SA. The fine-grained SA always gives us precise outcomes regarding text polarity, whereas positive or negative emotions can be obtained through emotion-detection SA. The aspect-based SA is capable of determining sentiments corresponding to different aspects for a single entity. Finally, the fourth significant SA kind provides us more profound concern related to the user’s intention. This type of SA is the hardest among all four types. There are multitudes of challenges associated with the process of finding ...