Automatic feature extraction based structure decomposition method for multi-classification

摘要: For years, researchers in neural network (NN) area have been carried out much productive research improving the generalization ability of NNs. In this paper, a novel design algorithm is presented for solving multi-class problems, structure decomposition based on Skeletonization (SDBSkeletonization), which to simplify NNs further. The proposed method decomposes complex problem into set two-class each can be regarded as an individual problem. After learning all these problems parallel with algorithm, we then integrate results final decision. addition, solves classification automatic feature extraction. This perspective gives broader range application our method. Our experimental Waveform and Handwritten Digits database demonstrate that SDBSkeletonization improves overall performance.