摘要
为提高销售预测准确率,提出一种基于改进XGBoost的销售预测方法。首先对销售量影响因素进行特征分析,通过改进灰色关联分析方法对训练数据进行降维处理;然后采用基于XGBoost算法的销售预测方法对降维后的特征数据进行监督训练;最后使用训练后的模型对销售情况进行预测评估。实验结果表明,基于灰色关联分析和XGBoost模型的销售预测方法正确率达到95%以上,比传统的经典预测方法提高35%以上,比XGBoost预测方法提高19.6%。基于灰色关联分析与XGBoost模型的销售预测方法不仅能有效处理海量数据,提高销售预测准确率,还能为制造企业实现产品精准投放提供决策依据。
A sales forecasting method based on improved XGBoost is proposed. Firstly,the influential factors that affect the sales volume are analyzed by characteristics,and the training data is reduced in dimension by improving the grey correlation analysis method.Then,the sales prediction method based on the XGBoost algorithm is used to supervise and train the feature data after dimensionality reduction. Finally,the trained model is used to predict and evaluate the sales situation. The experimental results show that the accuracy rate of sales forecasting method based on gray correlation analysis and XGBoost model reaches more than 95%,which is 35% higher than the traditional classic prediction method,and 19.6% higher than the XGBoost prediction method. The sales forecasting method based on gray correlation analysis and XGBoost model can not only effectively deal with massive data,improve the accuracy of sales forecasting,but also provide decision-making basis for manufacturing enterprises to achieve accurate product launch.
作者
张星
何利力
郑军红
ZHANG Xing;HE Li-li;ZHENG Jun-hong(Information Institute,Zhejiang Sci-Tech University,Hangzhou 310018,China)
出处
《软件导刊》
2020年第9期6-10,共5页
Software Guide
基金
国家重点研发计划子项目(2018YFB1700702)。
