A projection multi-objective SVM method for multi-class classification

Support Vector Machines (SVMs), originally proposed for classifications of two classes, have become a very popular technique in the machine learning field. For multi-class classifications, various single-objective models and multi-objective ones been proposed. However,in most models, neither different costs misclassifications nor users’ preferences were considered. This drawback has taken into account models.In these large hard second-order cone programs(SOCPs) constructed ane weakly Pareto-optimal solutions offered. In this paper, we propose Projected Multi-objective SVM (PM), which is that works higher dimensional space than object space. PM, can characterize associated solutions. Additionally, it significantly alleviates computational bottlenecks with numbers classes. From our experimental results, see PM outperforms SVMs (based on an all-together method, one-against-all method one-against-one method) other SVMs. Compared to SVMs, provides wider set options designed misclassifications, without sacrificing training time. methods, promises out-of-sample quality approximation Pareto frontier, considerable reduction burden.