Prediction of App Success Using Machine Learning Algorithm

Software development is based on implementation standards. In the case of selling and accepting software by customers, it has been a challenge to develop applications for marketplaces. App stores have features such as the number of downloads, comments, and ratings on ratings. From this, the difficulties were the fields (previously listed) and their way of analyzing the problem, thus resulting in characteristics that define the pattern of success in apps. Based on this scenario, this work aimed to create two inference engines from the SVM, Decision Tree, XGBoost and Random Forest algorithms and, with that, the features that determine the best correlation for the rating of the applications were investigated, besides to compute and evaluate regression metrics using a Google Play Store database. The work is structured as follows: in section II, the theoretical framework will be presented, where the main themes relevant to this work will be explained. Section III will discuss the materials and methods, where the step by step will be described according to the CRISP-DM to obtain knowledge of the database. In section IV, the results obtained with the execution of the algorithms will be structured and in section V, it will be the conclusion, which will be scored what was possible to understand with this research.