Enhancing Mobile App Recommendations With Crowd Sourced Educational Data Using Machine Learning And Deep Learning

Abstract

The rapid growth of mobile applications has created a need for intelligent recommendation systems that deliver personalized and context-aware suggestions to users. This study presents a hybrid recommendation framework titled “Enhancing Mobile App Recommendations With Crowdsourced Educational Data Using Machine Learning and Deep Learning.” The system leverages crowdsourced user interaction data, including demographic attributes, app usage behavior, ratings, and temporal features, to improve recommendation accuracy. The proposed model integrates multiple machine learning and deep learning techniques, including Random Forest, K-Nearest Neighbors (KNN), Artificial Neural Networks (ANN), and Long Short-Term Memory (LSTM) networks. Data preprocessing involves timestamp decomposition, categorical encoding using Label Encoding, and feature selection to ensure compatibility across models. The dataset is split into training and testing subsets to evaluate performance using standard metrics such as accuracy, precision, recall, and F1-score. A Flask-based web application is developed to provide an interactive interface where users input relevant parameters such as age, gender, app category, device type, and usage patterns. The system generates predictions from all trained models and outputs the recommended mobile application along with confidence scores. The inclusion of LSTM enables the model to capture temporal dependencies, while ensemble and instance-based methods enhance robustness and generalization. Experimental results demonstrate that combining traditional machine learning models with deep learning architectures improves recommendation effectiveness and adaptability. The proposed system provides a scalable and efficient solution for personalized mobile app recommendation using crowdsourced educational data, contributing to enhanced user engagement and decision-making.