Opinion Mining of Customer Reviews : Feature and Smiley Based Approach
I R Jayasekara, Wijayanayake W M J I · International Journal of Data Mining & Knowledge Management Process · 2016
With the rapid growth in ecommerce, reviews for popular products on the web have grown rapidly.Although these reviews are important for making decisions, it is difficult to read all the reviews.Automating the opinion mining process was identified as a solution for the problem.Although there are algorithms for opinion mining, an algorithm with better accuracy is needed.A feature and smiley based algorithm was developed which extracts product features from reviews based on feature frequency and generates an opinion summary based on product features.The algorithm was tested on downloaded customer reviews.The sentences were tagged, opinion words were extracted and opinion orientations were identified using semantic orientation of opinion words and smileys.Since the precision values for feature extraction and both precision and recall values for opinion orientation identification were improved by the new algorithm, it is more successful in opinion mining of customer reviews.