| 摘要 |
PROBLEM TO BE SOLVED: To provide a system for easily judging whether the name of a brand is similar or not. SOLUTION: A neural network is constructed by applying the feature parts of respective brands to be compared to an input layer and applying the similarity or non-similarity of both the brands to an output layer on the other hand. Input layer pattern information prepared concerning the brands to be compared is inputted to the input layer of the constructed neural network, and it is judged from the output value of the output layer whether the brands are similar or not. Besides, the similarity judgement based on a brand examination reference is processed according to a flowchart in a figure 29.
|
