Genetic Algorithm for Multilayer Radial Basis Function Networks Based on Generalized Inverse Matrix
-
摘要: 为了解决以往正则最小二乘法求权重向量时遇到的矩阵接近奇异而无法求逆的问题，采用广义逆矩阵的方法求多层径向基函数网络中各层的权重向量，并将这种方法引至多层径向基函数网络的遗传算法中。采用实函数逼近，混沌时间序列建模与预测等仿真实验对算法进行验证。结果表明，采用广义逆矩阵的方法要比正则最小二乘法在逼近精度上高1至2个数量级。Abstract: In order to solve the problem that maxtrice are nearly singular when using regular least squares method, a method of generalized inverse matrix was employed to obtain the weight vectors of each layer in the multilayer radial basis function network, which was introduced to the genetic algorithm for training multilayer radial basis function networks. By using real function approximation, chaotic time series modeling and forecasting simulation experiments, the algorithm was verified. The results show that the generalized inverse matrix method is much more better than regular least squares method on the approximation precision, which can be up to a 1 to 2 orders of magnitude.
-
-
[1] MOODYJ, DARKEN C. Fast learning in networks of locally-tuned processing units[J]. Neural Computation, 1989, 1(2):281-294.
[2] ZHANG J, JIANG H F, WU R L. Multi layer RBF network for real functions approximation and nonlinear regression[C/CD]//Scientific Research Publishing. Proceedings of 2011 International Symposium on Statistics & Management Science. Chongqing: Scientific Research Publishing, 2011.
[3] 吴瑞柳. 多层径向基函数网络的聚类学习方法及应用[D]. 马鞍山:安徽工业大学,2011. [4] 李建. 多层径向基函数网格的算法改进及其应用[D]. 马鞍山:安徽工业大学,2013. [5] 盛飞. 多层RBF网络的自适应遗传算法及其在实函数逼近中的应用[J]. 安徽工业大学学报(自然科学版),2013, 30(2):192-196. [6] 徐光鲁. 复合多层径向基函数网络及其在偏微分方程中的应用[J]. 安徽工业大学学报(自然科学版),2015, 32(1):76-80. [7] 余健,郭平. 基于RBF网络的金融时间序列预测[J]. 湖南工程学院学报(自然科学版),2007, 17(4):44-48. [8] LIU G R, GU Y T. 无网格法理论及程序设计[M]. 王建明,周学军,译. 济南:山东大学出版社,2007:259-264. [9] 张雄,刘岩. 无网格法[M]. 北京:清华大学出版社,2004:95-175. [10] 侯木舟,周耀琼. 二叉树期权定价方法的一种新推广[J]. 数学理论与应用,2006, 26(3):111-115. [11] 牟旷凝. 蒙特卡洛方法和拟蒙特卡洛方法在期权定价中应用的比较研究[J]. 科学技术与工程,2010, 10(8):1925-1929. [12] 蹇明,宜娜,张春晓. 期权定价模型的五点式混合差分方法[J]. 经济数学,2011, 28(4):66-70. [13] 王宋桂,杨振海. 广义逆矩阵及其应用[M]. 北京:北京工业大学出版社,1996.
计量
- 文章访问数: 45
- HTML全文浏览量: 0
- PDF下载量: 1
下载:
