Application of Random Forest Method to QSAR Model Building and Prediction of Toxicity
With growing environmental concern, a need for predicting the toxicity of compounds has emerged. Experimental assessment of toxicity can be costly, time consuming, and hazardous. Quantitative structure-activity relationships (QSARs) can be used to predict toxicity accurately based on experimentally known toxicities. QSARs modeling tools have traditionally been satisfied by the Statistics, Machine Learning methods. Considering the data dimension, descriptor selection, and prediction accuracy, Random Forest (RF) method was selected for the descriptor selection and model building in the present study.
random forest QSARs toxicity
Li Zhang Lin Yu Xinling Yang
PMDD Lab. , Department of Applied Chemistry, College of Science, China Agricultural University, Beij Department of Applied Mathematics, China Agricultural University, Beijing 100193, China
国际会议
北京
英文
381-382
2012-09-15（万方平台首次上网日期，不代表论文的发表时间）