Sentiment Analysis Using machines Learning Approaches of Twitter Data and Semantic Analysis

Abstract

The widespread use of the World Wide Web has ushered in a new way for people to share their feelings. It is also a medium with a wealth of knowledge where users can see other users' opinions, which are divided into various sentiment groups and are gradually becoming a key factor in decision- making. This paper contributes to the sentiment analysis for consumer review classification, which is useful for analysing information in the form of a large number of tweets with highly unstructured views that are either positive or negative, or somewhere in between. To do so, we first pre-processed the dataset, then extracted the adjectives from it that have some context, which is known as feature extraction. vector, then added the function vector list classification algorithms that use machine learning, such as: The content function is extracted using Naive Bayes, Maximum Entropy, and SVM, as well as the Semantic Orientation based WordNet, which extracts synonyms and similarity. Finally, we evaluated the classifier's output in terms of recall, precision, and accuracy.