| 摘要 |
PROBLEM TO BE SOLVED: To automatically acquire meaning knowledge (case frame pattern). SOLUTION: A dictionary part 3 stores corresponding information between a word and an attribute showing a class to which the word belongs. A case frame pattern learning part 4 inputs a case frame from a case frame storage part 1, refers to the dictionary part 3, substitutes the word in the case frame for the attribute, generates the attribute case frame, generates a depending graph with probability by learning based on the attribute case frame and stores the depending graph with probability in a case frame pattern storage part 2 as the case frame pattern. A natural language processing part 5 executes a natural language processing while the case frame pattern stored in the case frame pattern storage part 2 is refereed to. |
