Hybrid Naive Bayes with Pigeon Optimization for Classification of Opinion Mining
Twitter is a common way to express yourself and connect with others in the online world. Therefore, Twitter is measured a tremendous source of info for decision making and emotional analysis. When predicting the polarity of words, the analysis refers to a classification problem and then categorizes them into positive and negative emotions, in order to identify attitudes and opinions expressed in some way or form. Mental analysis through Twitter provides organizations with a quick and effective way to monitor public attitudes towards their brand, and directors. Recent research has focused on the different aspects and methods used to train sentiment classifiers for a set of Twitter data with different results. The main problems of the previous techniques are the accuracy of the classification, the sharpness of the data and the ridicule, since most of the tweets categorize and neutralize a high ratio of tweets. This study focuses on these issues and presents an algorithm for rating Twitter feed based on a hybrid approach. In our scheme is applies to different stages of pre-processing before placing the text in the classifier. The research notes show that the projected method has exceeded the previous limits and performs better compared to the corresponding measures.