论文阅读(2013-2)----a new collaborative filtering-based recommender system for manufacturing appstore: which applications would be useful to your busines?

  论文阅读(2013-2)----a new collaborative filtering-based recommender 
system for manufacturing appstore: which applications would be useful to your 
busines?

  这篇论文讲的东西并不深,讲的是appstore上的app个性化推荐问题,简单做个笔记。

简单介绍:

推荐系统可以降低没有卖任何app就离开的用户的概率、当用户买了某个app后,可以推荐配套的app、增加用户的忠诚度。

思路介绍:

Ui=(用户在频道1购买的app数目,用户在频道2购买的app数目,...用户在频道n购买的app数目)

US(用户相似度)=(Ui*Uj)/(||Ui||*||Uj||)

首先通过和你相似topK个用户购买了的app而你没有买的app作为候选推荐app,然后,就是对这些候选app的一个rank问题。

rank=候选app和自己购买app的共现对/自己购买app数。

结论:

个人觉得,思路清晰,相对简单。

论文阅读(2013-2)----a new collaborative filtering-based recommender system for manufacturing appstore: which applications would be useful to your busines?,布布扣,bubuko.com

论文阅读(2013-2)----a new collaborative filtering-based recommender system for manufacturing appstore: which applications would be useful to your busines?

上一篇:Android 开发Project中各个目录和文件的介绍


下一篇:Android TabHost TabWidget 去除黑线(底部下划线)