文化算法优化rbf神经网络及应用分析-optimization of rbf neural network by cultural algorithm and its application analysis.docxVIP

太原理牛 太原理牛 大学硕士研 究生学位论文 太 太原理工大学硕士研究生学位论文 THE APPLICATION AND RESEARCH OF RBFNEURALNET 明ORK OPTIMIZED BY CULTURAL ALGORITHMS ABSTRACT With the ever-accelerating pace of life，people are more eager to easily and quickly communicate with the computer directly. Voice is an ideal human-compute r interaction ，making it a focus of research in this area. Speech recognition technology apphcation carries out this mission for us. 1n recent years，speech recognition technology is stiU mainly based on the study of nonJinear systerns，which fetters the further development of this technology. The researchers reaJized that the introduction of nonlinear methods is imperative to achieve substantive progress. 1n the 1980s， NonJinear-theory such as artificial neural network used for solving nonlinear problems become a researcb hotspot. As a new and effective feedforward neuraI network，RBF gets increasing attention and applies in the 自eld of speech recognition. The topology parameters of RBF are determined dynamically during the training 10 make the networks adaptability better. However ，the performance of tbe network is very sensitive to the selection of the center of the basis function. As a new optimization ，cultural a1gorithms have better gIobaI search performance. YA-- YA - - This paper uses the cultural aJgorithms to determine the center of the basis function to improve the RBF neural networks performance ，and apply it in speech recognition system. The main contents are as follows: 1、Introduce a new culture algorithms 仕amework model with the global search performance. Since this aJgorithm is stiJJ in its infancy，its intrinsic mode has not formed. The model depends on the setting of specific issues. According to the need of this paper ，the components of cultural algorithms framework has been desined to solve the nonlinear unconstrained optimization problems. Cultural algorithms use the genetic algorithm as the populat