An assessment of data mining techniques reliability in predicting social media sentiments

Abstract

SNSs (Social networking Sites) are the most popular medium for global communications. Internet users have been increasing in tandem with evolving technologies where their expression about organization, events, personalities and many more other discussions can be found in online review sites or social networks media or blogs. SAs (Sentiment Analyses) are a part of ongoing researches using DMTs (data Mining Techniques). They are computational treatments of opinions, sentiments and subjectivity of text. SNSs generate voluminous data that can play an essential role in decision making of individuals, organizations and even governments. It impossible to scrutinize texts or sentiments expressed on SNSs where SAs provide polarities to the text and classify text into positive or negative categories where DMTs can be used for categorizations though they may result in different accuracy percentages. The main aim of this study is to examine categorizing of sentiments expressed in SNSs by DMTs and evaluating them in terms of accuracies or speeds of executions.