WEIGHT BASED PRE-PROCESSING ALGORITHM (WBPA) FOR DATA PRE-PROCESSING IN SENTIMENT ANALYSIS

Abstract

In the big data period, data is made ceaselessly or nearer to constant. Social media, like Twitter, makes a gigantic proportion of such data. Regardless, social media data are every now and again unstructured and testing to make due. Thusly, this paper proposes a convincing text data pre-processing technique Weight Based Pre-processing Algorithm (WBPA) to deal with yahoo finance data. Foster an algorithm that weights the feeling score similarly as weight of hashtag and cleaned text and foster an algorithm to weight the scores of the hashtag and cleaned text to secure the opinion. The results show that stemming technique performed best with respect to computational speed. Besides, the accuracy of the algorithm was attempted against actually organized sentiments and sentiments conveyed before text data pre-processing. To the extent that model execution, the proposed algorithm performed better with the higher accuracy perhaps, due to the unstructured idea of yahoo finance data.