| 摘要 |
PROBLEM TO BE SOLVED: To speedily search a data base by detecting all polisemic words in some data base character string during inquiring, rearranging them according to a found adjustment state and adjusting the adjustment cost, and repeating this rearraying and readjusting processes. SOLUTION: Polysemic words found in both an inquiry character string and a data base character string is confirmed and the costs for the respective polysemic words are generated. The appearance of the polysemic words in the inquiry character string is adjusted to the appearance of the polysemic words in each data base character string to numerically represent the similarity between the inquiry character string and data base character string. An edge part present in the adjustment solution is inspected, the inquiry character string is rearrayed to reduce the cost, and the adjustment and rearraying are repeated at a specific frequency or until no improvement is obtained in spite of the trouble of the adjustment. To confirm the data base character string which is the most similar to the inquiry character string, those processes are repeated for respective data base character strings.
|
