Primary Liver Cancer Early Diagnosing Model Based On Machine Learning

Primary liver cancer is a malignant tumor with high mortality.Its causes are complex and early symptoms are not obvious.The diagnosis method is usually based on the imaging data of the liver.If there is an abnormality,it is determined by biopsy.However,these methods have the disadvantages of invasiveness and radiation and are not friendly to patients.With the gradual maturity of medical informatization,the number of clinical data has exponentially increased.Use artificial intelligence-related methods to analyze the potential information of data has become a research hotspot in the medical field.Based on the machine learning method,this paper studies the tabular data in the clinical field and establishes a diagnostic model.The specific work is as follows:(1)Classical machine learning algorithms achieve an early diagnosis.This part contains two modules.The first is to combine the support vector machine with the differential evolution algorithm,the optimization algorithm is used to improve the selection method of kernel function parameter,and the cross-validation is introduced to prevent the model from overfitting;the second is to use the gradient decision boosting tree to perform features select and compare the performance difference between the model using all features and some features.Experiments show that the improved algorithm has higher performance with an accuracy of 0.9441;feature selection reduces the amount of calculation while ensuring the accuracy of prediction,which is 0.9541.(2)Deep learning methods achieve the classification and staging of liver cancer.First,the clinical test data is converted into image data in the form of two-dimensional character embedding;second,based on transfer learning,using a pre-trained model of the residual network,changing its final fully connected layer,fixing other layers,and fine-tuning parameters;then,train and verify models.Experiments show that changing the data format can improve the performance of the model,and its accuracy is close to 0.7,and the AUC value is above 0.8.The research results can help clinicians in the early warning of liver cancer and decision-making of classification and staging,which is conducive to the early and quick diagnosis of the disease,thereby improving the survival rate of patients.