Multi-class support vector machine with random inputs using non-deterministic programming problem

This paper applies the expected value to release the constraints from the probabilistic state.Then, the sample mean is used to estimate the expected value in the unknown population state.Also, the Monte Carlo method is implemented to simulate the data, and then the bootstrap resampling method is used to generate random samples to estimate the expected value.Moreover, the confusion matrix statistical indices and accuracy criterion are used to evaluate the proposed and competing models. Results and DiscussionUsing three kernel functions, linear, polynomial, and Gaussian, we investigate the accuracy values of the three models.In the simulation subsection, we examine the performance of the three proposed models, basic and Cramer-Singer, using the statistical indices of mean and standard deviation, and we compare the three models using the Friedman test statistic.Also, in the real example subsection, we evaluate these models in different parameter ranges for three real data sets of seeds, ecoli, and leukemia. ConclusionGiven that most of the data we encounter in the real world is not precise and certain and has inherent flaws due to measurement errors, this paper proposes a model that addresses this issue.The results show that the model based on random inputs is far better than the models based on definite and precise inputs.