基于数据挖掘的电商订单转化率的预测

Prediction of Conversion Rate of E-Commerce Orders Based on Data Mining

摘要: 电商营业增速趋于平缓，流量和移动互联网红利等带来的超高速增长基本结束，电商业务进入精细化运营阶段。不同于线下的运营模式，电商的订单量主要受营销策略的影响，且可以收集到详细的运营数据，具有极大的利用潜力。本文通过建模的方式，利用线性模型和非线性模型，尤其是机器学习模型，以转化率为目标进行回归预测。本文的研究意义及目的在于帮助平台型电商进行转化率的预测，以辅助根据不同阶段的企业目标制定不同的转化率目标，进行营销资源的分配。目前的研究多以理论研究为主，缺少对具体应用环节的分析，本文从数据和模型的角度，从特征选择到选择模型为逻辑进行展开，对营销策略的影响进行建模分析，并得出相应的结论，具有实践指导价值。

Abstract: The growth of e-commerce business tended to be flat and the ultra-high-speed growth brought by traffic and mobile Internet dividend was basically over. E-commerce business entered a re-fined operation phase. Different from the offline operation mode, the electricity supplier’s order quantity is mainly affected by the marketing strategy and can collect detailed operational data, which has great potential for utilization. In this paper, by way of modeling, linear regression model and non-linear model, that is machine learning model, are used to predict the conversion rate. The purpose of this paper is to help platform-based e-commerce providers to predict the conversion rate and to assist in setting different conversion rates according to the different stages of business objectives and allocating marketing resources. At present, most researches mainly focus on theoretical research, lacking of analysis of specific application links. From the perspective of data and models, this paper starts from the feature selection to the selection of models as the logic to carry out modeling analysis of the impact of marketing strategy and draws corresponding conclusions, with practical guidance value.