A novel class of layered feedforward neural network models for function approximation was proposed based on the principle of multi - dimensional discrete fourier transform 摘要利用多維離散傅立葉變換原理構(gòu)造新穎的神經(jīng)網(wǎng)絡(luò)模型用于函數(shù)逼近,網(wǎng)絡(luò)結(jié)構(gòu)為分層前向網(wǎng)絡(luò)。
Abstract : based on the strong learning ability and nonlinear function approximation capacity of multi - layer perceptrons ( mlps ) , a generating chaotic sequence model is proposed in this paper 文摘:應(yīng)用具有全局最優(yōu)的進(jìn)化規(guī)劃算法建立產(chǎn)生混沌序列的優(yōu)化神經(jīng)網(wǎng)絡(luò)模型。
In chapter two , the basic knowledge related to panel display drive technology and chromaticity is introduced . some terminology of functions approximation theory is presented 第二章介紹了有關(guān)平板顯示器驅(qū)動(dòng)技術(shù)的基本原理和顯示色度學(xué)相關(guān)的基礎(chǔ)知識(shí),同時(shí)給出函數(shù)逼近理論的基本概念。
Second , its generalization ability through the concrete function approximation example is analyzed and a method for obtaining better generalization ability through dynamic gauss width searching is presented 同時(shí),通過(guò)函數(shù)仿真的實(shí)例分析了該網(wǎng)絡(luò)的泛化能力,并給出了一種通過(guò)動(dòng)態(tài)自動(dòng)尋優(yōu)獲得該網(wǎng)絡(luò)較好泛化能力的方法。
Conditions that a class of sequence has convergent subsequence arc discussed in the paper , this sequence is important in function approximation . the gained conclusions are useful in some relative areas 摘要文章討論了在函數(shù)逼近論中有重要作用的一類(lèi)序列存在收斂子列的條件,文中所得結(jié)論在相關(guān)問(wèn)題的研究中有較重要的作用。
In addition , the paper makes relatively in - depth analyses on the function approximation theory of radial basis function neural networks and the stability of the adaptive controller based on radial basis function neural networks 此外,本文對(duì)徑向基函數(shù)神經(jīng)網(wǎng)絡(luò)的函數(shù)逼近理論以及基于徑向基函數(shù)神經(jīng)網(wǎng)絡(luò)的自適應(yīng)控制器的穩(wěn)定性作了較深入的分析。
In the course of the practicality , the professional knowledge and experiences , or the capability of nonlinear function approximation are adopted in fuzzy modeling , then genetic arithmetic is applied to optimize the network 實(shí)際過(guò)程中采用了專(zhuān)家知識(shí)和經(jīng)驗(yàn)進(jìn)行模糊建模,或利用神經(jīng)網(wǎng)絡(luò)的非線性函數(shù)逼近能力建模,然后應(yīng)用遺傳算法對(duì)網(wǎng)絡(luò)進(jìn)行優(yōu)化的技術(shù)。
Svm , developed from that theoretical architecture , is a highly adaptive method , which is applied in the areas of pattern recognition , regression estimation , function approximation and density estimation 在這一理論基礎(chǔ)上發(fā)展了一種新的通用學(xué)習(xí)方法一支撐向量機(jī)svm 。它是一種普遍適用的方法,已經(jīng)廣泛的用于模式識(shí)別、回歸估計(jì)、函數(shù)逼近、密度估計(jì)等方面。
We look at the problem of learning from examples as the problem of multivariate function approximation from sparse chosen data , and then consider the case in which the data are drawn , instead of chosen , according to a probability measure 并檢視稀疏精選值中多變量函數(shù)近似法等這些從實(shí)例學(xué)習(xí)法所發(fā)現(xiàn)的問(wèn)題,然后根據(jù)機(jī)率衡量,審思隨機(jī)獲得資料而非選定資料的案例。
Based on the result above , the open loop control strategy is bringed forward in the paper . we use the function approximation property of neural network to obtain the function through studying using neural network method . and then realize the position control 利用前向bp網(wǎng)絡(luò)能夠?qū)θ我夂瘮?shù)以任意精度逼近的特點(diǎn),對(duì)開(kāi)環(huán)控制策略中難以確定的函數(shù)進(jìn)行學(xué)習(xí),實(shí)現(xiàn)兩關(guān)節(jié)的任意位置控制。
The need for function approximations arises in many branches of applied mathematics, and computer science in particular. In general, a function approximation problem asks us to select a function among a well-defined class that closely matches ("approximates") a target function in a task-specific way.