基于随机森林模型的电动汽车目标客户销售策略研究

Sales Strategy for Electric Vehicles on Target Customers Based on Random Forest Model

为科学制定销售策略,提高电动汽车销售成功率,以2021年华数杯C题为例,针对电动汽车销售体验数据,采用箱型图检测筛选异常值并用相应均值进行替换,采用随机森林回归模型填补缺失值。在此基础上,以目标客户满意程度与客户特征为模型输入,以目标客户是否购车作为模型输出,建立随机森林模型,得到各影响因素的重要程度,并对潜在客户的购车意愿进行预测,准确度达97%。最后,通过变步长网格法对潜在客户满意度提升量进行调整,在潜在客户购买的前提下,优化选取出服务难度最小的销售策略。

To develop the scientific sales strategies and thus increase the sales of electric vehicles,Question C in the 2021 Huashu Cup Competition is taken as an example. Based on the data of sales experience of electric vehicles,the box graph is adopted to detect outliers and replace them with corresponding mean values.The random forest regression model is adopted to fill in the missing values. Then, with the satisfaction of target customers and customer characteristics as the input of model and whether target customers buy cars or not as the output of model,the random forest model is created to obtain the importance of each factor and to predict the purchase behavior of potential customers. The accuracy of 97 % is achieved. Finally,the variable-step grid method is used to make an adjustment to the improvement of satisfaction of potential customers. On the condition that potential customers buy the electric motors, the sales strategy with the least difficulty in service is optimized and selected.