Sentiment Insights Unveiling Text Emotions Through Machine Learning

Abstract

Sentiment analysis, a pivotal tool in understanding human emotions and opinions, is the focus of this research paper. The project, developed by a team of four members, aims to provide a web-based application for sentiment analysis with various features. The paper explores the significance of sentiment analysis and the motivation behind the project. The project encompasses several key features. Firstly, the sentiment analysis feature categorizes text into positive, negative, or neutral sentiments and displays the corresponding emoji for each result. This classification is achieved using the textblob.polarity library for prediction. Secondly, the emotional analysis feature delves deeper into sentiment analysis by categorizing sentiments into emotions such as happy, love, worry, hate, and sadness. This feature utilizes SVM and LSTM models and provides emoji's for each emotion. Additionally, it facilitates the analysis of tweets, allowing for the manual addition of tweets or their import from Twitter using an API (although the API is currently unavailable). Another significant feature of the project is the product reviews sentiment analysis, which includes a word cloud and sentiment analysis to classify words based on emotions. It also analyzes product reviews from customers for a list of products and utilizes data visualization to display negative and positive reviews. Furthermore, the project includes audio sentiment analysis, which enables users to utilize call recordings, convert them into transcripts, and conduct sentiment analysis. It uses a pie chart to illustrate whether the sentiment is positive or negative and highlights positive and negative lines in the transcript. This feature utilizes NLTK for transcript processing and sentiment analysis. Lastly, the project includes story sentiment analysis, which evaluates the sentiment of a story and character sentiment using a story listener. It uses a web API to fetch stories and provides features such as comparing two characters and the sentiment of the story at intervals. Data visualization is used to present all this information.