Machine Learning Based Mobile Sales Prediction System

The mobile phone market is characterized by intense competition, rapid technological evolution, and dynamic consumer preferences, making accurate sales forecasting a critical yet challenging task for manufacturers and retailers. This paper presents a machine learning-based system for predicting the sales performance of mobile phones. The system is developed by training and evaluating a suite of diverse machine learning algorithms, including Linear Regression, Random Forest, Gradient Boosting, and a simple Neural Network, on a dataset comprising various mobile phone attributes such as RAM, battery capacity, screen size, and price. The objective is to identify the model that most accurately predicts the sales volume or price range. Each model is trained on the same feature set and evaluated using standard regression metrics like Mean Absolute Error (MAE) and R-squared (R²) score. Our comparative analysis demonstrates that tree-based ensemble models, particularly Gradient Boosting, provide the most accurate predictions by effectively capturing the non-linear relationships between a phone’s features and its market performance. The resulting system serves as a valuable tool for strategic decision-making in pricing, marketing, and inventory management.