Abstract: Surgery time duration estimation is the premise and basis of operation room scheduling. In order to estimate the surgery duration time accurately and provide effective information for operation room scheduling, by using the cross validation method to optimize the support vector machine parameter, a cross validation and support vector machine estimation model (CVSVM) is built. In order to verify the performance of the model, the model is compared with the radial basis function (RBF) neural networks model, and a case of eye cornea transplant of a hospital is used to validate the model. Results show that, in comparison with the RBF model, the mean absolute percentage error of CVSVM is below 11% and the relative error is below 23%, confirming the effectiveness of the model and providing an effective method to estimate surgery time duration.

Key words: surgery time duration (CV), cross validation, support vector machine (SVM)