A Novel Machine Learning Approach for Sentiment Analysis Based on Adverb-Adjective-Noun-Verb (AANV) Combinations

The capability to study facts(data) about each living as well as non-living entity and derive conclusions(information) from those facts and then store them for future use and reference(knowledge), is an art which no other species has been gifted. This skill has been enriched over the time. With the advent of the internet, communicating across the globe has virtually been reduced to our palm. So, it is of utmost importance, to judicially use our vocabulary and grammar, to get the true feeling and sentiment across to the intended person/(s). Almost no research work has been done , as on date, based on adverb-adjective-noun-verb (AANV) combinations in sentiment analysis .We have proposed here for the first time , an AANV based sentiment analysis technique deploying linguistic analysis of adverbs, adjective, abstract noun and categorized verb, which has been a significant advancement from the previous research on this domain. We define a set of general axioms for opinion analysis to determine a functional value of the sentiment analysis. In this analysis, Entropy, Conditional Entropy and Information Gain are some of the well-defined concepts that have been applied to evaluate our proposed opinion analysis system.