Multi-feature fusion based drug side effect predicating method

Abstract

The invention discloses a multi-feature fusion based drug side effect predicating method. The method is characterized in that a multi-label integrated K proximity method is carried out to predicate the drug side effects based on four attributes of a drug; a large number of side effect data experiment results show that the method is high in predicating accuracy, and high in robustness; in addition, the side effects which can only be detected after certain drugs are marketed can be successfully predicated. Therefore, the method is applicable to drug side effect safety evaluation and patient clinical medication reference.

Description

technical field [0001] The present invention relates to bioinformatics, especially a method for predicting drug side effects based on multi-feature fusion. The method is based on four characteristics of drugs and uses a multi-label integrated K-nearest neighbor method to predict drug side effects. Background technique [0002] In recent years, the safety issues in the pharmaceutical field have aroused widespread attention at home and abroad. Researchers have begun to use pattern recognition, complex network theory and other methods to mine new potential drug side effects information and evaluate drug safety from massive drug-related data information. So far, researchers at home and abroad have roughly proposed four types of drug side effect prediction methods: methods based on statistics, methods based on molecular docking, methods based on text mining, and methods based on machine learning. [0003] In 2009, Fukuzaki et al. used the statistical similarity model based on th...

AI Technical Summary

Problems solved by technology

However, its features are used too much, and only biological characteristics include drug target pro

Images