Mobile Phone Price Range Prediction Based on Machine Learning Algorithms

Abstract

Mobile phones with different price points on the market have different parameter configurations, which provide a lot of information related to the price. Mobile phone prices are also the focus of attention of the mobile phone market, manufacturers and consumers. This study used machine learning technology to investigate the issues in mobile phone sales, mainly the prediction of mobile phone price range and the feature contribution of the price range. The data resource includes the release price of mobile phones in the European market from 2018 to 2021. The data augmentation technologies are also applied in this study. Among the three established machine learning models, the accuracy rate of the training set of the Catboost model exceeds 80%, which is the model with the highest accuracy rate. The accuracy rate of Bagging Classifier is slightly lower than that of CatBoost, but the prediction result is the most stable. Such a result conforms to the characteristics of the algorithm itself. The conclusion of the analysis of feature contribution of numerical data to accuracy rate is that rear camera pixels, number of cameras and screen size have negligible impact on model prediction accuracy, which means feature contribution is minimal. The research results are reasonable and can provide references for merchants to set prices and for users to purchase mobile phones.