Rule-based Learning In Medical Literature Automatic Indexing System

Traditional Chinese Medicine (TCM) is a medicine of Chinese, which has several thousands of year's history. The prosperity of Chinese proved that TCM has great live and value in its existence. Nowadays, TCM's modernization is the key area of research in medicine and other subjects, like chemistry and pharmacy. With the development of informatics, the data about TCM began more than ever. A great deal of data make some users confused.Nowadays, The literature database of TCM has been established. There are almost several ten thousand literature every year. How to make use of computer technology to complete some editing task, such as indexing, the extraction of key word automaticly and semiautomaticly, and reduce the artificial uncertainty and errors in literature editing, reduce the expense of man and thing, improve the efficiency and quality of literature classification is very important.The paper provides a system using rule-based learning method of automatic indexing for medical literature. The system has a relatively high performance in subject indexing problem, which involves the procedure of combination of medical headings and subheadings. Also the paper gives automatic indexing experiment in medical literature, which shows the system is much better than the former one. The main contributions are following:1 To analyze whisk algorithm in information extraction, and make correspond modification of whisk for Chinese medical literature data. And the modified whisk is made as the rule learning algorithm of the automatic indexing system.2 To develop a system of word management. Because subject word database and entrance word database change endlessly, the change bring big effect for the automatic indexing result. So we develop a system of word management, responsible for the updating of Mesh word, the association of subject word and entrance word, the analyse and statistics of updating word.3 To advise developing rule set by rule learning, and develop the rule set by modified whisk algorithm, also make a automatic indexing experiment in 2001 year literature, the result shows the system is a processing system.