Rational curves and rational surfaces , which are a class of important approximation functions , are extensive applied in cad / cam 有理曲線和曲面作為一類重要的逼近函數(shù),在計(jì)算機(jī)輔助設(shè)計(jì)與制造中有著廣泛的應(yīng)用。
This paper approximates non - linear function with wavelet neural network , whose ability of approximation function is discussed in theory 本文將小波神經(jīng)網(wǎng)絡(luò)用于逼近非線性函數(shù),并從理論上討論了它逼近函數(shù)的能力。
By applying the result to approximation function / which is " essentially localized " in time - frequency , we obtain good approximation effect 作為應(yīng)用,用該非調(diào)和小波函數(shù)逼近一個(gè)時(shí)頻局部化的函數(shù),收到較好的效果。
In the first part , this thesis emphasizes to introduce the approximation functions and inducement of three filters . also this thesis introduces the reason of choosing the cauer approximation 著重介紹了三種濾波器的近似函數(shù)以及其推導(dǎo)過(guò)程,并介紹了為什么要選擇橢圓近似。