摘要
介绍了紧支撑小波神经网络（ＣＳＷＮＮ） 的理论和算法， 将其运用于钢的冷弯性能判别， 判别准确率达到１００ ％ 。与ＢＰ算法的计算值和文献中所采用的主成分分析（ＰＣＡ） 比较， ＣＳＷＮＮ 的判别能力优于ＢＰ及文献中采用的主成分分析。
The theory and algorithm of compactly supported wavelet neural network (CSWNN) were introduced and applied to the discrimination of cool bending natures of steels. The discrimination accuracy is 100% with CSWNN. Predicted results indicated that CSWNN was better than the back\|propagation (BP) algorithm and the principal comporent analysis (PCA).
出处
《计算机与应用化学》
CAS
CSCD
1999年第6期447-450,458,共5页
Computers and Applied Chemistry
基金
国家自然科学基金!29775001
关键词
紧支撑
小波神经网络
钢
冷弯性能
判别
CSWNN
Compactly supported wavelet neural network, Wavelet analysis, Discrimination, Cool bending nature of steel
