An Improved DDAGSVM by Using the Value of Separation Margin

摘要

Decision directed acyclic graph support vector machine (DDAGSVM) is an effective approach to solve multi-class problem, but it has to solve the problem of how to choose the structure of the graph and minimizing the classification error that might be accumulated at the final classification process. In order to improve the generalization ability of DDAGSVM, and minimizing the classification error that might be accumulated at the final classification process, the efficient method is studied in this paper. Based on the idea that to maintain high generalization ability, the most-easily separated classes should be separated firstly during the formation of DDAG, and by using the value of margin as the value of separability measure, an improved algorithm for DDAGSVM is proposed. With the given algorithm, the accumulation of classification error can be reduced efficiently. By using the value of separation margin as the value of separability measure in the formation of DDAG, the training time increasing can be avoided. Classification experiments prove the effectiveness of the given algorithm.