福建电脑2026，Vol.42Issue(1)：7-14,8.DOI:10.16707/j.cnki.fjpc.2026.01.002
钻石价格预测模型比较及优化路径
Comparison and Optimization Path of Diamond Price Prediction Models
摘要
Abstract
To accurately predict diamond prices,this paper constructs four regression models:multiple linear regression,random forest,GBDT,and SVM,and systematically compares their prediction performance and optimization effects.The research first performs data preprocessing,divides the dataset,and standardizes the features.Subsequently,through three stages:initial evaluation,K-fold cross-validation,and grid search hyperparameter optimization,using RMSE,MAE,R²,and MAPE as evaluation metrics,combined with visualization methods,the model performance is analyzed.The results show that the optimized random forest performs best in terms of R²(0.9775)and MAPE(6.92%);GBDT shows significant optimization effects,with RMSE reduced by about 27%;multiple linear regression consistently performs well in various indicators but has the worst overall performance;SVM shows performance degradation after optimization.The research results can provide a reference for diamond pricing and regression model selection.关键词
钻石价格预测/回归分析/模型评估/超参数优化Key words
Diamond Price Prediction/Regression Analysis/Model Evaluation/Hyperparameter Optimization分类
信息技术与安全科学引用本文复制引用
葛艳娜,陈春娣,理艳荣,朱士玲..钻石价格预测模型比较及优化路径[J].福建电脑,2026,42(1):7-14,8.基金项目
本文得到广东省高等教育教学改革项目《基于OBE理念的大数据专业项目导向教学模式研究与实践》(No.2023JXGG05)、广州商学院校级科研课题:数智融合优秀课程项目(No.XJYXKC202539)资助. （No.2023JXGG05）
