Mining Conflict Semantic from Drug Dataset for Detecting Drug Conflict

Abstract

The detecting of drug interactions hiding in the massive drug data, especially the conflict (i.e., some drugs react with each other) detecting, plays an important role in the medical information field. This kind of conflict detecting can not only relieve the cognitive burden for doctors, but also help some people (e.g., physicians and patients etc.) avoid the risk of reactions among drugs in some extend. This paper presents a Drug Conflict Detecting (DCD) algorithm to rapidly find reactions among several drugs according to the user's query requirements. First, the user dictionary and waste words base are built according the data feature of medical data sources to effectively extract drug term including component and interaction terms. Then, all conflict semantics are mined to establish conflict knowledge base based on the results of drug term extraction. Finally, the DCD algorithm is proposed to provide rapid detection of drug conflict. The experimental results show that the proposed algorithm has high accuracy. It can effectively and rapidly implement the drug conflict detecting.