Prediction of Antimicrobial Resistance Based on Random Forest Algorithms
Download as PDF
DOI: 10.25236/cmph.2019.026
Author(s)
Xiujuan Xie, Xiangju Li, Yu Sheng, Bing Gu
Corresponding Author
Xiujuan Xie
Abstract
Most hospitals use traditional bacterial culture methods to detect bacterial resistance, which has a long cycle and delays doctors'understanding of patients' antimicrobial resistance, and brings challenges and difficulties to clinical drug use. Therefore, this paper proposes to apply random forest algorithm of data mining classification method in antimicrobial drug resistance detection. Using a large number of historical data of bacterial susceptibility testing in a third-class A hospital as the original data set, the classification model of antimicrobial resistance is obtained through pretreatment, model training and model evaluation, and then the new strains can be predicted. In addition, some representative data sets are selected and compared with the traditional decision tree C4.5 algorithm. The experimental results show that the random forest algorithm has better prediction effect and performance, and has certain practical application value.
Keywords
Drug Resistance Prediction; Antibiotics; Random Forest Algorithm;Data Mining