Clinical drug-drug adverse reaction detection method based on label propagation algorithm

Abstract

The invention relates to a clinical drug-drug adverse reaction detection method based on a label propagation algorithm. The drug-drug adverse reaction is further detected based on novel similarity based on a given drug sample set and label initialization reconstitution of a label propagation mode. The method comprises the steps of firstly, filtering drug characteristics by adopting a CHI method, selecting the characteristic that the information amount is large; secondly, constructing a new sample similarity according to the sample label similarity and the sample similarity adjusted by the laplace operator; thirdly, establishing the initialization information of the unknown label sample based on the known label sample information; finally, detecting the adverse reaction of the drug throughlabel propagation. According to the method, the drug similarity calculation mode and the label propagation mode are reconstructed, so that the similarity between drugs is more accurate, the label propagation mode is smoother, and the detection of drug-drug adverse reactions in the clinical stage can be effectively improved.

Description

technical field [0001] The invention relates to the field of drug safety detection, in particular to a clinical drug-adverse drug reaction detection method based on a label propagation algorithm. Background technique [0002] In traditional drug safety detection methods, such as frequency method (reporting ratio ratio method (PRR), reporting odds ratio method (ROR) and comprehensive standard method (MHRA)) and Bayesian method (Bayesian confidence proliferating neural network (BCPNN) and Multivariate Gamma-Poisson Distribution Reduction Method (MGPS)) and other methods are for the detection of drugs with adverse reactions in the market. In real life, it is necessary to test the drugs before they go on the market to prevent some unsafe drugs from appearing on the market. After taking these unsafe drugs, they may cause other diseases or cause the death of patients. In recent years, with the popularity of big data, big data methods have also been used in the medical field to de...

AI Technical Summary

Problems solved by technology

However, in addition to the shortcomings of the similarity method, this method also has shortcomings in the selection of sample data feature information and the lack of unlabeled sample label initialization methods.

Images