This paper aims to investigate the problems on classification of merchandise information for home living. Four daily newspapers were used as information sources. Information was abstracted from these sources, and arranged into short sentences using data-base system. Eight students of Yokohama National University separately selected key words from 23 items of merchandise codes. Key words were put into key word fields of the data-base system. The deviation of key words was checked. Merchandise codes can be separated into five groups, from the point of correlation of 23 items of the key words. The information, which has difficulty in classifying, can be separated into two types ; one type involves a few key words in information, and another type is related to all sorts of merchandise.
View full abstract