Support vector machine (SVM) modeling is one of the most popular machine learning approaches in chemoinformatics and drug design. The influence of training set composition and size on predictions currently is an underinvestigated issue in SVM modeling. In this study, we have derived SVM classification and ranking models for a variety of compound activity classes under systematic variation of the number of positive and negative training examples. With increasing numbers of negative training compounds, SVM classification calculations became increasingly accurate and stable. However, this was only the case if a required threshold of positive training examples was also reached. In addition, consideration of class weights and optimization of cost factors substantially aided in balancing the calculations for increasing numbers of negative training examples. Taken together, the results of our analysis have practical implications for SVM learning and the prediction of active compounds. For all compound classes under study, top recall performance and independence of compound recall of training set composition wasachieved when 250–500 active and 500–1000 randomly selectedinactive training instances were used. However, as long as ∼50known active compounds were available for training, increasing numbers of 500–1000randomly selected negative training examples significantly improvedmodel performance and gave very similar results for different trainingsets.
展开▼
