Authors: Fujun Gao
Addresses: Zibo Vocational Institute, Zibo, Shandong, 255314, China
Abstract: This paper first constructs two sets of datasets to demonstrate the effectiveness of the proposed method, one dataset consists of all human protein data, and the other is composed of human G protein-coupled receptor data, which accounts for a high proportion of drug targets. It extracts the corresponding primary structure, polypeptide characteristics and basic physicochemical properties of each protein in the dataset, feature selection is used to reduce the learning burden of classifier as the feature space of training classifier. Then the data are preprocessed and the optimal classifier is constructed by adjusting the parameters of the model. Datasets are classified by SVM classifier and Adaboost-SVM classifier respectively in the experimental construction and analysis part, analysed and compared the experimental results of two classifiers applied to two sets of datasets before and after data preprocessing, the classification results of the two groups were verified each other to increase the reliability of the classification results. The experimental results verify the effectiveness of the proposed method. At the same time, it shows that the method proposed in this paper can effectively predict drug targets, and provide a preliminary reference for drug research and development workers.
Keywords: direct push type; support vector machine; SVM; predictive drugs; target quantification; structure-activity analysis.
International Journal of Innovative Computing and Applications, 2020 Vol.11 No.2/3, pp.109 - 114
Received: 12 Mar 2019
Accepted: 29 Apr 2019
Published online: 30 Apr 2020 *