Emote_X: An AI-Driven Framework for Real-Time Twitter Sentiment Classification Using Machine Learning

Abstract

Social media has become one of the ways people share their thoughts, feelings and responses to what is happening in the world. Twitter, which is now called X is an example. It creates a constant flow of short messages. These messages contain clues about how people feel in areas like business, health care, politics, entertainment and when there is a disaster. Tweets are very messy. They have abbreviations, hashtags, emojis, slang and spelling that's not consistent. This makes it hard to look at them manually. It takes a time and is not always correct.

To fix this problem we created Emote_X. It is a system that does sentiment analysis on Twitter in time. It uses machine learning and natural language processing. The system gets tweets through the Twitter API. It. Prepares the text to remove the noise. Then it turns the text into numbers using TF-IDF vectorization. After that three learning methods. Logistic Regression, Support Vector Machine and Naïve Bayes. Are trained to put tweets into negative or neutral groups. Once they are classified the results are shown using dashboards. This helps users see trends and patterns in sentiment easily.

Our tests show that Emote_X gives results while using little computing power. This makes it good for real-time use. Emote_X is not for research. It is also useful in areas, like keeping track of how a brand's seen, looking at customer comments watching public opinion and making decisions based on data. Together Emote_X gives a simple, growing and easy way to do automatic sentiment analysis. It shows how machine learning and NLP can take the mess of media text and turn it into real value.