602.txt

This complete matlab for neural network

发信人: fervvac (高远), 信区: DataMining
标  题: Re: 倒排索引如何实现？
发信站: 南京大学小百合站 (Wed Oct 30 13:37:20 2002), 站内信件

Thanks for your introduction. 

I still don't know the exact problem your are facing. Anyway, in the field of
IR, they define a set of keywords w_i,. Given a set of document D_i, each D_i
contains possibly multiple keywords. An inverted index built on the vocalbury
W and document set D is , conceptually, an array of linked lists. Each element
in the array stands for  a keyword, and the associated list represents all
occurance of the keywords in the documents, i.e., d_i -> d_j -> ..., ->d_m.

Efficient algorithms exist to build such inverted index , which uses memory
carefully.

Not sure if your problem can be reduced to that one.

【 在 mywingnd (飞飞) 的大作中提到: 】
: 
: 【 在 fervvac 的大作中提到: 】
: : What's your *** matrix?
: : And what's your queries?
: : Efficient algorithms to construct inverted index for text data exist. But ..
: : sure they are what you wnt.
: : 【 在 mywingnd (飞飞) 的大作中提到: 】
: 用户项矩阵是关于用户对产品的评价信息的数据集，主要用于电子商务网站的推荐系统，
: 在实现推荐系统的技术中，有一种是协同过滤技术，它就考虑通过对与当前用户有相似兴
: 趣的邻居用户的考虑，来预测或推荐该当前用户对某一产品的兴趣。因此希望对这一数据
: 集建立倒排索引


--
※ 来源:．南京大学小百合站 bbs.nju.edu.cn．[FROM: 143.89.41.4]