App Success Classification Using Machine Learning Models
Biplab Thapa Magar, Subin Mali, Eman Abdelfattah · 2021
Millions of mobile apps are downloaded and uploaded to the Google Play Store platform every day. Various features of an app, features that are either internal to or external to it, can play a significant role in determining whether the app will be successful or not. Using a dataset containing data extracted from the Google Play Store, this paper uses various features external to an app in order to make classifications on the app's success. This research therefore deals with a classification problem by classifying apps based on various degrees of success. Various classification models are created and compared with an aim to get a better understanding of classifying an app's success. The models used are Logistic Regression, K-Nearest Neighbors, Stochastic Gradient Descent, Decision Trees, Random Forest and Support Vector Machine. The models are compared based on their performance measures and run times. Lastly, the paper presents the classification model with the best classification performance while also exploring the relationship between the number of target categories and the performance of each classification model.