Method for constructing postoperative recovery state prediction model of surgical patient

The invention discloses a surgical patient postoperative recovery state prediction model construction method. The method comprises the following steps: 1, converting a time sequence variable into a trend variable; 2, determining a postoperative recovery state variable of each index by using the trend variable and the normal range of the index; 3, utilizing a support vector machine (SVM) algorithm to classify the postoperative recovery state variables of the patient according to the postoperative recovery state variables converted from all the time sequence variables and other variables except for the number of bed-leaving activities and the time length of the bed-leaving activities for 3 days after the operation in the non-time sequence variables; 4, by taking the postoperative recovery state classification as dependent variables and taking preoperative variables, postoperative patient behavior variables, postoperative bed-leaving activity times and postoperative bed-leaving activity duration cut-off 3 days i