Research on RS-WNN Integrated Modeling
Wavelet Neural Network (WNN) integration modeling based on Rough Set (RS) is studied. An integration modeling algorithm named RS-WNN, which first introduces a heuristic attribute reduction recursion algorithm to determine the optimum decision attributes and then conducts WNN modeling, is proposed. This method is adopted to more effectively eliminate the redundant attributes, lower the structure complexity of WNN, which reduce the time of training and improve the generalization ability of WNN. The result of the experiment shows this method is superior and efficient.
Paul P. Lin and Chunliang Zhang
Z. C. Xu et al., "Research on RS-WNN Integrated Modeling", Applied Mechanics and Materials, Vols. 105-107, pp. 2169-2173, 2012