PAC分析とテキストマイニング手法を併用した学習動機の構造分析

抄録

In this study, we try to use “PAC & Mining” analysis for the survey of the learning motivation of three Japanese adults who are studying Portuguese. The procedure of the method is as follows. The first stage procedure is to carry out a clustering based on the similarity of words among associated items by text mining methods. Secondly, the clustering based on the similarity of a subject’s intuitive evaluation on associated items by PAC analysis is compared with the results of the clustering by text mining methods. Finally, the structure of the attitude and image of a subject are studied from these two viewpoints. As a result of the analysis of the learning motivation, it is found that the key words which are extracted by text mining methods and characterize each cluster are dispersed into many clusters or into one or few concentrated clusters in the cluster structure of PAC analysis. The concentration and dispersion of these key words show what kind of learning motivation the learner is related to. It is suggested that “PAC & Mining” analysis offers a viewpoint that cannot be obtained by the traditional PAC analysis.