Toward development of a health behavior promotion model: An analysis of healthcare data by using PLSA

Abstract

This paper considers an analysis of healthcare data by using PLSA toward development of a health behavior promotion model. In order to apply PLSA to healthcare data, we adjust the weights of factors of healthcare data based on the patterns of health guidance and analyze the effectiveness of adjusting weights. The experimental result of clustering indicated that the clustering when adjusting weights brought about better clus-tering performance regarding the patterns of health guidance than clustering where the weights were fixed, by the evaluation measures: Precision, Recall and F-measure.Moreover, we found that increasing the degree of the weights enhances the clustering performance particularly in terms of Recall.