abstract |
The invention discloses a method for constructing a diabetes risk model based on a scoring system, which belongs to the field of big data and machine learning. Feature extraction, obtain feature set, perform feature screening on the obtained feature set, obtain the optimal feature subset after screening, bin each feature in the obtained optimal feature subset, calculate the WOE of each box and map In the data, use logistic regression modeling, use scoring module to make diabetes risk score card; adopt the risk model constructed by the method of the present invention, can obtain more refined and quantified diabetes risk score, using this model, It is conducive to guiding individuals to conduct more appropriate personal health management, including more reasonable eating habits, exercise habits and other living habits, and is conducive to commercial insurance institutions to make more accurate and objective commercial underwriting decisions. |