Feature Selection Based Naïve Bayes Algorithm for Twitter Sentiment Analysis

A intriguing areas of study now is Twitter sentiment investigation. It fuses the data mining methods used to create such systems with natural language processing techniques. The majority of currently available methods for analyzing Twitter sentiment do not do well when presented with messages that are both brief and ambiguous, since this is the only kind of information they take into account. Better classification outcomes are guaranteed if the right collection of features is used to identify emotion in online textual material. The computational difficulty of doing optimal feature selection drives the need for the development of creative approaches to enhancing classifier performance. In this work, an effective method for analyzing Twitter user sentiment was presented. A machine learning model was developed by the suggested method to identify good and negative tweets. Characteristics are extracted after pre-processing, and it's possible that include irrelevant features can lower classification accuracy. For this reason, the model implemented the Vortex Search Algorithm (VSA) to choose the best characteristics and discard the rest. The final tweets categorization is done using the Naive Bayes (NB) algorithm. A total of four Twitter datasets are used for the studies, each measuring a unique set of parameters and made accessible to the public. Marketing, detecting political polarization, and product reviews are just a few of the many areas that may benefit from the suggested system's ability to gauge user sentiment based on tweets.