Leveraging ANFIS and Evolutionary Algorithms to Predict Mobile App Success on Google Play

The rapid expansion of the mobile phone market has driven the growth of mobile applications, making the global app industry highly competitive and lucrative. As developers strive to maintain revenue and market position, predicting the success of mobile apps before their release has become increasingly important. This research focuses on the Google Play Store, the largest app marketplace worldwide, where metrics like app installations and user ratings serve as key indicators of market performance in the absence of direct revenue data. To address this challenge, we developed a predictive model using the Adaptive Neuro-Fuzzy Inference System (ANFIS), enhanced by optimization techniques such as Differential Evolution (DE), Particle Swarm Optimization (PSO), and Genetic Algorithm (GA). The study utilized a dataset of 10,841 Android applications from the Google Play Store, key input features include Category, Size, Type, and Content Rating, while the output is defined as the product of user Rating and Installations to reflect app popularity. The predictive model demonstrated that Differential Evolution (DE) achieved the highest performance, followed closely by Particle Swarm Optimization (PSO), with the Genetic Algorithm (GA) also significantly enhancing the baseline ANFIS model. These findings highlight the importance of dataset-driven feature selection and optimization techniques in improving predictive accuracy. This research provides developers with a robust framework for anticipating app success, offering valuable insights in the competitive and rapidly evolving mobile app ecosystem.