A Study Of Auxiliary Decision For Liver Transplantation Based On Machine Learning

Liver transplantation is one of the most effective treatments for acute liver failure,liver cirrhosis and even liver cancer.The prediction of postoperative complications is a significant work in modern medicine.It will be helpful for liver transplantation to accurately predict postoperative complications.Although many machine learning algorithms can well deal with the problem of prediction and classification,these methods have the problems of low prediction accuracy,low accuracy,and low recall when they are used in medical data sets with small samples and large feature space.Scientific analysis and processing of liver transplantation data to obtain effective information is of great significance in assisting doctors to further study the survival rate of liver transplantation.In this paper,for the decision-making problem of liver transplantation surgery,based on the liver transplantation surgery data provided by the cooperative hospital,to help doctors improve the accuracy of postoperative complications diagnosis,a liver transplantation complications prediction method based on transfer component analysis and support vector machine is proposed.to solve the problem of the small amount of medical data,this paper uses the method of migration component analysis and progressive alignment heterogeneous migration learning to map and reduce the dimension of feature space and maps the source domain and target domain to the same reproducing kernel Hilbert space to realize the adaptive edge distribution.Also besides,this paper uses PCA to compare the different SVM(support vector machine),KNN(k-nearest neighbor),and XGBoost(extreme gradient)The experimental results show that compared with the traditional PCA dimensionality reduction method and the progressive alignment heterogeneous migration learning dimensionality reduction method,the proposed dimensionality reduction method greatly improves the accuracy and F1 value of the complication prediction.