NLP Based Sentiment Analysis Using Particle Swarm Optimization with Logistic Regression for Digital Education

Identification and categorization of user sentiments from a text passage is known as sentiment analysis into several sentiments. For example, using feelings like happiness, sadness, anger, or positive, negative, or neutral to gauge how the user feels about a certain topic or item. These days, a variety of websites and applications collectively referred to as online educational platforms allow users to obtain a higher degree of education across a range of areas. For this reason, digital education provides a platform for the dissemination of knowledge. Among the most active areas within NLP, web mining, and text mining research is sentiment analysis. It is significant in a variety of subjects, including as the social sciences and management sciences, where evaluating the efficacy of learning technology depends heavily on student input. Due to the proliferation of educational institutions, many students have become enthralled with online learning portals that provide cost-free courses. Every year, a huge number of students enrol in these expansive online courses and then revisit their opinions regarding the course material and level of instruction. Additionally, share ideas on blogs on how to raise the standard of instruction by expressing both good and negative opinions. However, an assortment of ML techniques are available to assess these ratings and ascertain which virtual education platform is the most user-friendly and optimal. This study examined natural language processing (NLP)-based sentiment analysis and social media evaluations that were manually gathered from educational applications. To categorise the reviewer’s comment for digital education, Particle Swarm Optimisation with Logistic Regression (PSOLR) outperformed all other existing Naïve Bayes, Decision Tree, and Logistic Regression algorithms in terms of accuracy (93.75%).