ANALYZING SENTIMENTS ON X USING MACHINE LEARNING

Abstract

Nowadays social networking platform has become a quick source of information and a part of daily life routine. Among all the emerging technologies X formerly known as twitter is the most popular platform to share information whether it is from daily life, industry, economy, politics or related to any other field people are tweeting every minute. There is a huge opportunity for data analysts to predict the trends by analyzing the data. This huge number of tweets attracts the attention of data scientists for analyzing the sentiments. By utilizing the sentiment analysis technique, they can classify the data into different categories like positive, negative, irrelevant or neutral and get important information from this. This study will show us how we can use social media for analyzing sentiments to get useful information. We will categorize the reviews of different users into different category using different algorithm. Categorizing the reviews help us to give information about different topics that how people feels about specific things and what are their opinions related to product through analysis we can improve strategies, advertisement, and known people interests. X formally twitter categories sentiments datasets sourced from Kaggle, which comprises user reviews about games, sentiments will be labeled as positive, negative, neutral and irrelevant. Python will employ model classification utilizing Logistic Regression, K-Nearest Neighbors, Decision Tree, Random Forest, and Ensemble Learning algorithms. Additionally, Python will be utilized for sentiment categorization into positive, negative, neutral, and irrelevant categories. The success rates of the classification algorithms will be compared to find out their respective performance levels. Experimental results revealed that ensemble learning exhibited the highest accuracy performance, achieving a remarkable accuracy of 91% compared to other methodologies.