According to the statistical results of the Ministry of Health and Welfare in 2012, circulatory diseases accounted for three of the top ten causes of death in Taiwan and became an important focus of medical care. Therefore, this study pays special attention to the correlation between cardiac catheterization and medical expenses. Taking the inpatients of DRG124 and DRG125 as the objects, using artificial intelligence methods, aiming at the clinical characteristics, demographic characteristics, medical history and other factors at the time of admission, a prediction module is established to know the medical expenses and payment areas of patients in advance. The research statistical methods include descriptive statistics, inferential statistics, multiple linear regression model establishment, training and evaluation, and random forest model establishment and training. The stochastic gradient descent method is used to optimize the multiple linear regression model, and the performance of the model is evaluated by MSE and R squared values. In the random forest model, multi-class classification, training and adjusting parameters are carried out, and finally the performance of the model is evaluated. The expected results hope to improve the accuracy of the prediction model to more than70%, and help medical institutions identify loss-making patients in advance. Through intelligent prediction, accurate allocation of resources and formulation of treatment plans, medical benefits will be improved and the medical system will be developed in a more accurate and efficient direction.

Keywords: Taiwan Diagnosis-Related Groups, Artificial Intelligence, Prediction Module