Hepatitis C Risk Prediction Based on Adaboost

Abstract

Hepatitis C is one of the major public health threats. The incidence of liver cirrhosis in 20 years after infection is about 20%, and the annual incidence of hepatocellular carcinoma is 2% - 4%, which is extremely harmful to the health and life of patients. However, people's understanding of hepatitis C is not comprehensive, and only 1 percent of hepatitis C patients worldwide have received effective treatment. At the same time, the early symptoms of hepatitis C are not obvious, and the differences between acute and chronic hepatitis C are large, leading many people to miss the best time for treatment. Therefore, reasonable prediction and classification of hepatitis C at an early stage can provide the most accurate medical guidance for patients and people with related symptoms. Machine learning is widely used in the prediction and classification of diseases in various medical fields, and its maturity has also been widely verified. In this paper, several types of machine learning models represented by decision trees are constructed in Python language to learn and predict the data provided by Ainshams University, and the accuracy rate is 72%. Finally, the data of the data set is analyzed, and relevant suggestions for preventing hepatitis C and in the treatment process are given.