AUTOMATIC BAYESIAN METHODS IN TEXT RETRIEVAL

点击翻译按钮获取中文摘要

A new model based on the Bayesian approach has been developed which has interesting connections with the vector models of G. Salton. Theoretical details have been worked out. Ideally documents must be indexed by the "real" objects that they refer to and these real objects become nodes in a system of multiple hierarchies called a specificity network. Each hierarchy is produced by a specificity operator and results in a tree of objects starting at the root with the most general and moving to greater specificity as one progresses towards the leaves. The objects which populate nodes are represented by textual terms or phrases. There may be many representations of any single object. As a first step in investigating this model a relational database has been constructed in Visual FoxPro that will contain the phrases extracted from each document and allow one to edit the phrases in a convenient manner. The initial set of phrases has also been extracted from one of the test sets of humanly judged MEDLINE records and we expect soon to embark on the task of editing this database. This task is the first stage of the project. The next stage will involve the examination of all the phrases in the database to see which pairs of phrases are related to each other in the set. When this has been completed we will have a new kind of indexing for the documents and it is our purpose to study retrieval with this indexing to see if it is free from some of the ambiguities that plague ordinary indexing and its associated retrieval.