ABSTRACT
Support vector machines (SVMs) are typically used for text classification, spam filtering, image recognition, colors classification, and currently for handwritten digit recognition currently used in post office automation. An example of SVMs use in the chemical sciences and engineering can highlight the research approach and the solutions that can be achieved by SVMs. SVMs furthermore can treat both linear and non-linear processes at once, so the problem of underfitting can be reduced, and if the number of parameters in the organizing hypotheses can be limited, there hopefully can be found a happy medium between under- and overfitting. Because SVMs can provide classification in a very high-dimensional hyperspace, there is in principle no limit to the number of factor segregations possible, and the SVM appears ideal for the highly complex computational chemistry of industrial processes that have limited experimental data.
