Font Size: a A A

Research On Enterprise Information Retrieval

Posted on:2007-01-06Degree:MasterType:Thesis
Country:ChinaCandidate:Y H ChenFull Text:PDF
GTID:2178360185968217Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
Enterprise information retrieval is a new topic in the field of information retrieval. It is designed to provide quick analyze for enterprises to deal with the data in their intranet, realize an effective content management, and then get a high productivity. Enterprise IR is a synthetically topic, which gives the chance to combine various IR technologies, e.g. auto extraction, web search, information filter, name-entity recognition etc. Also, enterprise IR is an original topic with enlightens. With analyzing to the enterprise network circumstance and enterprise data, new IR models and methods can be explored, and that's will promote the development of IR.Enterprise IR defines sub-issues by the requirement of enterprise, which means the target of retrieval. TREC2005 raised three sub-issues: known-item search, discussion search and enterprise search~2.This thesis focuses on the expert retrieval and the work which has been finished are as follows:1. Research on the three IR models and investigate their difference.2. Design a broken-line model as an overall model of expert retrieval, which allows using web search technology to support expert retrieval.3. Analyzing the relationship between key words and document, on the analogy of this, we discussed an expert weight formula and a method for rank of expert.4. We discussed a heuristic method to resolve the homonymy in name...
Keywords/Search Tags:enterprise information retrieval, expert search, homonymy, name-entity recognition
PDF Full Text Request
Related items