Efficient social media sentiment analysis using confidence interval-based classification of online product brands

Abstract

This paper presents an efficient method for categorizing the sentiments of Internet users, with a focus on social media users, using a confidence interval to estimate the reliability of sentiment predictions. The classification is based on the sentiments expressed in their posts, which are divided into positive, negative, and neutral categories. The paper presents an analysis table that analyzes sentiments and opinions about online product brands. The process includes two steps: 1) analyzing sentiments from text data using machine learning techniques, and 2) describing a five-step sentiment and opinion classification process that includes data collection, preprocessing, algorithm application, validation, and visualization. The proposed solution is implemented using Python, along with the scikit-learn, NumPy, pandas, and Dash libraries, and leverages the use of confidence intervals to assess the accuracy and reliability of the sentiment analysis model.