Application Of Machine Learning And MiRNA In Distinguishing Liver Cancer,Other Liver Injury And Normal People

In order to explore the feasibility and effectiveness of micro-RNA(miRNA)in the early diagnosis of liver injury,we utilized machine learning method with miRNA biomarkers to diagnose liver injury.This study is mainly divided into two parts.In the first part,through literature reading,the candidate miRNA biomarkers for liver injury were collected,and 13 micro-RNAs were obtained,namely: miR-21,miR-203 a,miR-214,miR-205,miR-451,miR-126,miR-192,miR-200 a,miR-221,miR-122,miR-16,miR-195,miR-206.In the second part,through three batches of data collection,six micro-RNAs were finally identified,namely miR-200 a,miR-221,miR-122,miR-16,miR-195,miR-206 as the final liver injury diagnosis organism Markers,and verified the classification effect of these six micro-RNA in machine learning classification.Among them,the first batch and the second batch of data are processed by machine learning methods.We use the 6 micro-RNAs with the best performance among the 13 micro-RNAs as the reference for the third batch of data collection.The third batch of data is machine learning The res?lts obtained from the processing res?lts further verify the effectiveness of these six micro-RNA selections.In addition,the four machine learning classification methods used in this study are TREE-J48 under the decision tree algorithm,Logistic under the logistic regression algorithm,Bayes net under the Bayesian algorithm and Na?ve bayes and LIBSVM.The optimal data processing method TREE-48 is obtained by analyzing different machine learning methods of data.