基于神经网络的口碑商家推荐

Recommender System on Koubei Platform Based on Neural Networks

介绍推荐系统在口碑平台的应用。在提取用户特性和时间特性以及商家特性的基础上,使用BP神经网络给用户推荐商家,同时对比BP神经网络中单隐层和双隐层的效果区别。对有线上记录的用户使用协同过滤进行推荐,对于无线上记录的新用户由于没有淘宝购物记录,无法判断其爱好,所以从商家特性进行分析并推荐给新用户。

Focuses on recommender system on Koubei platform. Based on the extraction of user characteristics, time characteristics and merchant characteristics, BP neural network is used to recommend users to the merchants. Then the effect difference between single hidden layer and double hidden layer in BP neural network is compared. Mainly uses the collaborative filtering to recommend users on the cable. For new us-ers recorded on the wireless, because there is no Taobao shopping record, so the recommendation is based on merchant characteristics.