| 摘要 |
PROBLEM TO BE SOLVED: To provide a data analyzer that enables the creation of long-term health prediction models using medical data having a deficit that makes it difficult to create long-term health prediction models as is.SOLUTION: An age bracket classifying unit 2 classifies word-frequency-type medical data obtained by evaluating health states of a plurality of respective subjects in one or more age brackets determined for each subject according to the age brackets. A complementing unit 4 compares medical data classified according to the age brackets between adjacent age brackets, and complements deficient medical data on a subject on the basis of medical data of another subject that has a resemblance to the subject and that exists on medical data on the adjacent age bracket. A prediction unit 5 creates prediction models for health states of subjects of a series of age brackets by applying clustering considering a time series trend aligning each age bracket to medical data for the complemented series of age brackets.SELECTED DRAWING: Figure 3 |
