App Assessment with Three Phase Evidence System using Sentiment Analysis

The rapid increase in mobile industry results in large number of mobile apps arising in the market which can be downloaded by either paying or free of cost. The selection of an app for a category will be based on rating, review and ranking by the user. The current system to detect the app’s genuineness is based on any one of the parameter that takes time as it doesn’t consider the other two which fail to correlate the results. If app’s evaluation on any one of the parameter gives good results, then there is no need for other kinds of parameters. For this reason, a novel method is proposed to collect, review, rate, and rank an app which will be evaluated independently. In case of review, each review of an app undergoes sentiment analysis using Word2Vec model that predicts words closer to the target word and classify them as either positive or negative. The rating of an app is considered by setting a threshold value for evaluation. The ranking pattern of an app is analyzed which classified under rising, recession and maintenance phase. The results of the above three parameters are aggregated which gives the evidence to determine whether the app is trustworthy or not.