抄録
Product concept sentences express effectively the characteristic images of products to communicate with the user's images. In this study, it was found first that users couldn't recognize such images from the existent concept sentences. We noticed that the characteristic image of each product, expressed by concept sentences, changes according to the noun after the adjective, even if the adjective is the same word. Therefore, we have developed an analysis method of their characteristic images from text typed-data such as concept sentences, which is called the ‘Product image analysis method.’ In the procedure of this analysis method, “adjective＋noun” units as image words are collected and arranged. Next, the units are classified, paying attention to the meanings of the nouns in the units. In this way, the classifiers understand the elements of product images expressed by the units. By this procedure, 8 categories with characteristic images of cars have been extracted from their concept sentences as samples. Further, based on these results, we have obtained the findings about the composition technique of concept sentences. Finally, the practical application of the ‘Product image analysis method’ has been confirmed from these obtained results.
