A Hybrid Classifier Approach to Analyze Public Sentiments in Social Domain

In recent times, social media has garnered the attention of researchers worldwide. The rationale behind this phenomenon can be ascribed to the extensive pool of data that is accessible as a result of the active engagement of users on said platforms. The present study advances a new approach for conducting sentiment analysis on Twitter data through the utilization of a hybrid algorithm. The process of analysing the public sentiment towards a specific subject matter is a multifaceted undertaking that encompasses various components such as pre-processing, score computation, and classification algorithms. The presented paper introduces an innovative approach that incorporates the impact of additional tweets in the computation of scores. In addition to this, pre-processing takes into consideration grammatical errors and the influence of the location of the tweet's origin. A novel algorithm that combines K-Nearest Neighbors (KNN) and Naïve Bayes techniques has been proposed to overcome the limitations of previous algorithms in dealing with data sets with high dimensionality.