| 摘要 |
A value of check data for each check item is not determined with a set of information which is about whether or not the health state of a test subject is normal or whether or not the value exceeds a boundary value. Provided is a method for determining a health state using a unique feature for each check item. The method for determining the health state comprises a model construction process and a health determination process. The model construction process divides an area for the value of the check data per check item into a normal area, a boundary area, and an ideal area. The normal area and the ideal area are divided into a boundary value predetermined in an international agency or an institute. The boundary area approaches the boundary value. When the boundary area is set, a plurality of determination candidate models is determined according to the value of the data included in the boundary area. One determination candidate model is selected from the determination candidate models by using the data of the test subject of past health diagnosis. The health determination process determines whether or not the value of the check data of the test subject for some check item is normal by using a determination model. [Reference numerals] (AA) Start model construction processing;(BB) End;(S110) Read data (learning data) for model construction;(S112) Process area setting for a normal area, a boundary area, and an ideal area which are predetermined per item;(S114) Write a plurality of determination candidate models according to whether or not a value of the data included in the boundary area is normal;(S116) Verify a determination candidate model;(S118) Determine a determination model |
