A Recommendation System prototype for www.dianping.com

• data contains the dianping.com data file
• getdat_master.py is response to crawling the reviews from dianping.com
• file_process.py preprocesses the raw data, including cleaning up
• file_process_str2id.py is response to transfering item and user to id
• recommendations.py contains a series of recommendations algorithm, the recommned behaviour happened here. Based on Slope One and Collaborative Filtering
• DianpingRecommender.java Mahout based version, A/B test with python recommend engine

一个基于大众点评网的点评数据编写的推荐系统原型，可以为指定用户推荐火锅店或者为指定火锅店推荐潜在客户

• data目录包含从大众点评网获取的用户评论数据
• getdat_master.py负责数据的抓取，采用多进程同步抓取的方式
• file_process.py负责对原始数据的预处理和清洗
• file_process_str2id.py负责把商户名称和用户名称转换为id表示的形式
• recommendations.py是进行实际推荐的文件，推荐算法基于Slope One和协同过滤
• DianPingRecommender.java使用Mahout推荐，与py版本进行A/B test