ENHANCING KNOWLEDGE DISCOVERY FROM MULTIPLE DATA SETS USING MULTIPLE SUPPORT VECTOR MACHINES

摘要

A system and method for enhancing knowledge discovery from data using multiplelearning machines in general and multiple support vector machines inparticular. Training data for a learning machine is pre-processed in order toadd meaning thereto. Pre-processing data may involve transforming the datapoints and/or expanding the data points. By adding meaning to the data, thelearning machine is provided with a greater amount of information forprocessing. With regard to support vector machines in particular, the greaterthe amount of information that is processed, the better generalizations aboutthe data that may be derived. Multiple support vector machines, eachcomprising distinct kernels, are trained with the pre-processed training dataand are tested with test data that is pre-processed in the same manner. Thetest outputs from multiple support vector machines are compared in order todetermine which of the test outputs if any represents a optimal solution.Selection of one or more kernels may be adjusted and one or more supportvector machines may be retrained and retested. Optimal solutions based ondistinct input data sets may be combined to form a new input data set to beinput into one or more additional support vector machine.