Could Mobile Applications' Success be Increased via MachineLearning and Business Intelligence Methods?
No Thumbnail Available
Date
2020
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The number of applications developed for mobile platforms is increasing as developer interest is largely directed towards mobileplatforms. In addition to the IOS platform, Google Play Store, one of the main platforms where developed mobile applications arepublished, also has a lot of developer interest, especially because it is open source. But there is no platform that developers can benefitfrom for elements such as the success that the developed application can provide or what features it should have. In this study, thisproblem was addressed. Accordingly, it is aimed to estimate and classify success according to the characteristics of the developedapplication. In addition, the evaluation of the developed application within the scope of business intelligence according to thecharacteristics of the previously developed applications is one of the main points of the study. Within the scope of the research, DecisionTree Regressor (DTC), Random Forest Regressor (RFR), K-Neighbors Regressor (KNN) and AdaBoost Regressor (ABR) were usedfor application rating estimates and the accuracy of the metrics were tested with R square score (R2), Mean Square Error (MSE) andRoot Mean square Error (RMSE). Estimates for classification Random forest classification (RFC) decision tree classification (DTC),the K-Neighbors Classification (KNC), Classification MLP (MLP), AdaBoost Classification (ABC) and naïve Bayes (GNB) has beentested with the algorithms used and the accuracy of confusion matrix metrics. In this context, DTR with 80.73% and RFR with 82.89%gave the best results for rating estimation, DTC with 86.08% and RFC algorithms with 89.83% gave the best results for successclassification. All predictions made with machine learning management in the scope of the study are dynamically shown in the webinterface using the Flask framework. Therefore, a platform was created where developers could get Decision Support with businessintelligence and the resulting results were analyzed and transferred into the work. In this way, mobile application developers will beable to see their shortcomings, if any, and have a prediction in terms of success.