To provide a research gene product data classification system for efficiently carrying out research target retrieval by classifying respective data such as gene or compound/medicine and disease according to working, effects, and symptoms or the like, and for removing noise data, a data group whose number of times of document excerpt is large, and data which do not belong to any classification in an early stage.

Gene information, compound/medicine information, disease information and document information associated with the respective information are stored so as to be associated with each other in a data base, and when any of the gene, compound/medicine and disease from the data base is used as a keyword, the other information is scored in the order of high relevance, and ranking is prepared. Then, second ranking data are prepared from the first ranking data, and the second ranking information is arranged for each of the first ranking information. The similarity of the first ranking data is calculated by using the second ranking data, and the first ranking data are grouped according to the similarity.

COPYRIGHT: (C)2004,JPO