Discriminant Analysis Based on Incremental Pruning in LS-SVM

LS-SVM(least squares support vector machine) is a widely applicable and useful machine learning technique for discriminant analysis. LS-SVM can be a good substitute for statistical discrimination, but computational difficulties are still remained for the solution to the linear system of LS-SVM, particularly in case of a large data set. We propose a discriminant method based on the incremental pruning at each time of a new input data point coming. With a new input data point, the pruned support vectors are newly modified, maintaining the predetermined number of pruned support vectors. The modification is continued until the last input data point. Numerical studies are performed to compare proposed method with batch pruning method via two real data sets.