基于神经网络的电子商务平台重复购买客户预测

The Prediction to Repeat Purchase in E-commerce Platform Based on Neural Network

研究电子商务平台中重复购买客户预测模型,构造出多层神经网络。原始数据编码采用词嵌入的方式,对子模块的特征向量设计出来之后设计顶层重复购买预测模型和算法。通过客户的行为日志对潜在重复购买客户进行预测具有重要意义。该预测模型适用于电子商务平台的重复购买预测,并取得了较好的效果。

This paper studies the prediction model of repeat purchase in e-commerce platform,constructs a mufti-layer neural network.The original feature data is encoded by word embedding method,and after designing the feature vectors of the sub-modules,the design to the top-level repeat purchase prediction model and algorithm is finished.It is of great practical significance to predict potential repeat purchase customers by using customer behavior log data.This prediction model is suitable for the prediction of repeated purchase of e-commerce platform,and has achieved good results.