An improvement of information retrieval system using word location information KwangYoung Kim, DuSeok Jin, YoonSoo Choi, JinSuk Kim, Jerry Seo, YoungKyon Suh Korea Institute of Science and Technology Information Group for Intelligent Information system {kykim, dsjin, armian, jinsuk, jerryseo, yksuh}@kisti.re.kr Abstract Today its difficult to effective searching the internet because its a huge collection of documents. Currently most people search documents by just typing a few keywords. But this is very import keywords for searching documents. This paper suggests that the word location information is an import element for searching relevant documents. This paper suggests that using word location information of a user query can exactly improve the result of information retrieval system. This paper makes directly experiments on retrieval system with various weight methods of word location information. 1. Introduction With the rapid expansion and popularity of the World Wide Web and the internet, it has made huge index systems like Google, Yahoo and so on. These index systems are very good. But most people have a feeling that it is hard to find relevant documents. Generally most commercial information retrieval engines adopt Boolean query as a query type. A user query like Boolean type is useful to information retrieval engines [1]. Today most information retrieval systems based on a user query like Boolean type. A user query is only information which a user wants to find documents. So a user query is the most important element to find relevant documents. Most information retrieval systems use to judge whether this document is a relevant match through a user query. We must consider the best way to find the best match document through a user query. This paper suggests the word location information of a user query is a very import element to find relevant documents. The weight of word location information is a very import element to decide relevant documents from searching result. This paper makes experiments how word location information affects the searching result. This paper will makes up the indexing files of word location information for this experiment and makes experiments with various weight methods of word location information. 2. Structure indexing file Generally to find relevant documents have importantly used the document frequency of indexing word importantly because it affects search result pages. The document frequency of information retrieval system is generally used to judge whether this document is a relevant match or not. This paper has posting files 1 of the document frequency information. Location information of indexing word importantly improved the precision of searching result with document frequency [2]. This paper additionally consists of indexing files with word location information and makes new experiments with these. New indexing files consist of the posting and the word location information files like Figure 1. The indexing files of posting information have information of document frequency. The indexing files of word location information have real word location information. The indexing files of word location information specially consist of the page numbers and the word numbers like figure 1. 1 index files