| Summary: | 碩士 === 朝陽科技大學 === 資訊管理系碩士班 === 91 === As more and more electronic documents are generated, the management of structured documents and the efficiency of the retrieval system become significantly essential. With popular inverted file structure, still a large number of documents will be searched and examined when queried with multiple index terms; that is, any document contains any one of the index terms will be checked. An improved new index structure Term Inheritance Structure (TIS), which can reduce the search time when queried with multiple index terms, is proposed. In addition, a new algorithm of query-document similarity is defined to more precisely measure the relativeness of retrieved documents based on user’s index terms and term weights. Although it takes time and space to construct the data structure of the TIS, it only needs to be implemented once. Thereafter it may save a lot of search and retrieval time when querying frequently from a large document base.
|
|---|
