Weighted variable kernel support vector machine classifier for metabolomics data analysis

Metabolomics data from modern analytical instruments have become commonly more and more complex, which brings a lot of challenges to existing statistical modeling. Thus there is a need to develop new statistically efficient methods for mining the underlying metabolite information hidden in metabolomics. In this study, we provide a new strategy weighted variable kernel coupled with the support vector machine (SVM), which is termed as the WVKSVM approach. The WVKSVM approach by modifying the kernel matrix provides a feasible way to differentiate between the true and noise variables. Finally, examples are given specifically for modifying a Gaussian kernel. Compared with some popular classification methods such as Random forest (RF) and the normal SVM, the results show that WVKSVM has better prediction ability and improve the performance of SVM classifier.