They can map complicated non - linear functions at infinite precision . in the paper , bp networks are trained as a classifier of the condenser fault . the test shows validity of the classifier is high 本文針對凝汽器故障非線性的特點和神經(jīng)網(wǎng)絡所具有的很強的學習能力、適應能力和魯棒性的優(yōu)點,用神經(jīng)網(wǎng)絡對凝汽器故障分類,在matlab軟件中的實驗表明該方法十分有效。
Our results show that the optimal value with a parameter is a monotone decreasing sectionally continuous linear function g about parameter a , and the membership function of fuzzy objective sets are a linear function , written as c 經(jīng)過研究發(fā)現(xiàn),其最優(yōu)值是隨著參數(shù)減少而逐漸減少的分段連續(xù)的線性函數(shù)g 。另外,模糊目標集的隸屬函數(shù)也是一個線性函數(shù),記為c 。
The rolling and the horizontal motion of ship are coupled intensely . meanwhile , the coupling motion is often presented as non - linear function according to experiment results and the ship is usually disturbed by sea wave when it cruises 船舶的橫搖運動與水平面運動是強烈耦合的,而這種相互耦合的運動根據(jù)試驗數(shù)據(jù)往往被擬合為非線性函數(shù)的形式,同時船舶在海上運動又會受到海浪的干擾作用。
A general approach toward robust optimal design is proposed which is based on tolerance model and orthogonal experimental method , this method can deal with implicit and high non - linear functions and great variances in design variables 摘要結合容差模型和正交實驗法的各自特點,提出了一種較為實用的穩(wěn)健優(yōu)化方法,該方法可以處理設計函數(shù)為隱函數(shù)或高度非線性函數(shù),設計變量的容差可容許大些。
Because compared with traditional bp nn - based ds , rbf network can reach a willful approximation to non - linear function , solving built - in discipline problem of system . also , a good learning speed can be reached . feature extraction is the key problem in intrusion detection 因為與傳統(tǒng)的基于bp神經(jīng)網(wǎng)絡的入侵檢測系統(tǒng)相比, rbf網(wǎng)絡可以對非線形函數(shù)進行任意逼近,解決系統(tǒng)本身的內(nèi)在規(guī)律問題,而且速度也比較快。
Wheel company organizational structure designed linear function system based general manager , deputy general manager , secondary manager ( assistant manager ii ) , on - site engineers , leaders of class and group , operators , a total of six levels of management 本文的研究對象具有典型的國有大中型企業(yè)的特征。因此,本文對我國國有大中型企業(yè),特別是汽車零部件企業(yè)的組織結構優(yōu)化設計,提高企業(yè)的核心競爭力具有積極的借鑒意義。
Traditional finite element calculation of saturated seepage always takes permeability coefficient as a fixed value , actually the permeability coefficient of unsaturated soil appears to be in a non - linear function relation , and has close relation with volume moisture content and pore water pressure 摘要傳統(tǒng)的有限元飽和滲流計算往往把滲透系數(shù)取作定值,而實際上非飽和土滲透系數(shù)呈非線性函數(shù)關系,與體積含水率、孔隙水壓力密切相關。
This paper puts forward bp improvable algorithm and exponent prediction pattern depending on neural network " s approach ability to non - linear function . comparing with traditional prediction algorithm , bp algorithm is verified to be feasible and accurate on exponent prediction 本文借助神經(jīng)網(wǎng)絡對非線性函數(shù)的逼近能力,提出了bp算法的改進型算法及基于bp算法的指數(shù)預測模型,通過對比傳統(tǒng)預測算法,證實改進后bp算法用于證券預測的可行性及準確性。
Several simplified first principle models are used to improve performances of robustness and generalization . the data - driven model , which is either linear function or nonlinear function such as neural network , is adopted to compensate the non - modeling part of the simplified first principle model 多個簡化機理模型用來提高模型的魯棒性和外推能力;非機理模型用來補償簡化機理模型的未建模部分,該部分可以是線性模型,也可以是諸如神經(jīng)網(wǎng)絡的非線性模型。
The standard form of each function and the method of a non - linear function changed into linear one are shown . the forecast models and synthetic evaluation models for the land use are built by unified regression statistical analysis models . the reliability for forecast result and forecast precision are given 本文概述了一元回歸統(tǒng)計分析模型的幾種常見類型,給出了各種函數(shù)的曲線標準形式和將非線性問題線性化的方法;建立了土地利用的一元回歸分析的預報模型,并對預報結果及預報精度給出置信度分析。