RULE BASED APPROACH FOR DETECTION OF MOBILE RANKING FRAUD

Abstract

The technology advancement mobile industry is growing rapidly at a rapid pace and millions of apps continue to be developed every day. These apps are available from Google Play and Apple's stores, although there are other stores. Companies are trying to get noticed by users as they market their apps. They manipulate apps using R3 fraud methods such as Rating, Rating, and Reviews. These R3 features will make the app more popular than apps with lower R3 features. To make apps last in the market, a fraud mechanism could be implemented on top of the R3 features. A model is developed to detect fraud using the rating, review, and rating of mobile apps. This study will compare the model to existing models like KNN classification, Rule-based approach, Hidden Markov Model. Ensemble decision tree is our proposed model. It has proven to be more prominent than other models with accuracy greater than 85%, and thus can detect fraud over apps with the greatest prominence.