Predicting Smartphone Prices Based on Key Features Using Random Forest and Gradient Boosting Algorithms in a Data Mining Framework

Abstract

This study aims to predict smartphone prices using machine learning models, specifically Random Forest and Gradient Boosting algorithms, based on various smartphone features such as internal memory, RAM, processor speed, battery capacity, and camera specifications. The dataset, consisting of 980 smartphones available in India, was preprocessed to handle missing values and categorical variables, ensuring it was ready for model training. The models were evaluated using Mean Squared Error (MSE) and R-squared (R²) scores, with Gradient Boosting outperforming Random Forest in terms of predictive accuracy. Key findings from the feature importance analysis revealed that internal memory, RAM, and processor speed were the most influential features in determining smartphone prices. The results indicate that machine learning models, particularly tree-based algorithms, are effective tools for predicting smartphone prices based on hardware specifications. This study has practical implications for businesses and consumers, as it provides insights into the factors influencing smartphone prices, helping businesses optimize pricing strategies and assisting consumers in making more informed purchasing decisions. Future research could explore deep learning models and incorporate additional features, such as market demand and consumer sentiment, to improve prediction accuracy.