hidden layer 中文意思是什麼
hidden layer
解釋
隱蔽層-
We deduce frondose algorithm of three layers bp neural networks which is used in common, and discuss several important issues in designing neural networks which is used to forecast, for example, number of hidden layer, nerve cell number of hidden layer, epoch of learning, embryonic power value, decision of node number about input and outputo at the same time, this dissertation sums up things that conventional bp algorithm is improved on considering disadvantages of it
3推導了常用的三層bp神經網路具體演算法，討論了實際預測應用中神經網路設計方面的幾個重要問題，如隱層數、隱層神經元數、訓練次數、初始權值、輸入節點數以及輸出節點數的確定。同時，針對傳統bp演算法存在的各種各樣的缺點，文中綜述了對其改進的情況。 -
In addition, many other problems also exist in hardware neural network, including error problem, learning mode, parallel architecture, and also neural network inner linking problem, hidden layer and the realization of the multiplicator and etc. for instance, error problem : hardware neural network employs the limited precision, and will inevitably bring limited precision error
另外，硬體實現神經網路還存在誤差問題，學習方式，并行結構等方面的問題，還有神經網路內部的連接問題，隱層及乘法器的實現等等。如誤差問題，硬體實現神經網路使用的是有限精度，不可避免的會產生有限精度誤差，選取合適的精度，才能既適合空間的要求，又避免對網路的實現產生一定的影響。 -
The main purpose of this paper is to investigate the application of the neuron network for the daily exchange rate forecasting. generalized cross validation is introduced to determine the number of nodes of the hidden layer, several well known time series forecasting methods are also compared with the nn method in this paper
討論了人工神經網路在金融匯率預報中的應用。其中介紹了廣義交互驗證generalized cross validation法如何應用於確定神經網路中隱層的個數，並用實例說明了該方法甚至對復雜的非線性函數也可以得到很好的逼近。 -
In the second layer, k - nearest neighbor algorithm is introduced to ascertain searching scope firstly, and then the nerve cell function ' s parameter in hidden layers begin to be evolved in this scope. the least - square is also introduced to calculate connection power between hidden layer and output layer
其中在第二級演化中，先用最小鄰聚法確定搜索空間，然後再在此空廣西大學頎十論文i 13f神經網路在ect圖像重注中的應用研穴間內進行演化，其中用最小二乘法來確定從隱層到輸出層的連接權值。 -
In course of training the network, the structure of the network has been analysed. using the nonlinear theory, the node of the hidden layer is ascertained. by comparing this method is very efficient
（ 5 ）在網路訓練過程中，對神經網路結構進行了分析，建立了計算輸出和理想輸出關系非線性方程組，依據非線性方程理論闡述設計變量、樣本數量和輸出層單元數量的關系，確定了隱層神經元的數量，經比較該方法很有效。 -
The number of the hidden layers of mul - tilayer perceptrons ( mlps ) is analyzed, and three - layer perceptrons neural network is adopted ; by analyzing the mechanism of the neural cells in hidden layer, a method for combining genetic algorithm and bp algorithm to optimize the design of the neural networks is presented, and it solves the defects of getting into infinitesimal locally and low convergence efficiently, it can also solve the problem that it can usually obtain nearly global optimization solution within shorter time through using genetic algorithm method lonely ; several examples validate that this algorithm can simplify the neural networks effectively, and it makes the neural networks solve the practical problem of fault diagnosis more effectively
對多層感知器隱層數進行了分析，確定採用三層感知器神經網路;通過對隱層神經元作用機理的分析，引入了遺傳演算法與bp演算法相結合以優化設計神經網路的方法，有效地解決了bp演算法收斂速度慢和易陷入局部極小的弱點，還可以解決單獨利用遺傳演算法往往只能在短時間內尋找到接近全局最優的近優解的問題;通過實例驗證了這種演算法能夠有效地簡化神經網路，使神經網路更加有效地解決實際的故障診斷問題。 -
The key feature of the proposal approach is using feedback linearization method to design a single hidden layer artificial neural network whose weights is renewed online to augment parametric uncertainty and unmodeled dynamics
其特點是採用隱含反饋線性化方法，設計單隱層神經網路在線更新權值，自適應補償參數誤差和未建模動態。 -
Based on the aerodynamics, control, structural dynamics model of smart rotor in frequency domain deduced and the determination for the number of neurons in hidden layer, the neuro - emulator using multiple independent miso neural networks with its deduced matrix expression for the smart rotor is set up. the rate of training is improved by introducing the orthogonal selection applying for smart rotor to the selection of training cases in neural modeling
試驗結果驗證了該方法的可行性，在建立了帶有主動控制后緣附翼的智能旋翼系統氣動-控制-結構動力學數學模型的基礎上，提出了適用於智能旋翼建模的多神經網路並聯型式的頻域模型，並推導出其矩陣表達式，探討了隱含層神經元數的確定方法。 -
Experiment showed that the network with 6 nodes in each hidden layer has the best - forecast ability. the bp network was developed by training, b )
將反演得到的氣象站點相對濕度資料，拓展為模式網格點的濕度資料，從而與數值預報模式相匹配。 -
The bp neural network is applied to the recognition of wear particles, and a bp neural network sorting system expected to recognize severe wear particle, cutting wear particle, normal wear particle and fatigue wear particle is designed and trained. 6. the function of the neural network ' s hidden layer is analyzed
將神經網路應用於磨粒識別，設計磨粒分類器，在網路學習中運用改進的bp模型，識別嚴重滑動磨損磨粒、切削磨粒、正常磨損磨粒和疲勞點蝕磨粒，隨機選取50個樣本對分類器進行訓練。 -
Firstly, influence factors of generalization of neural network are presented in this thesis, in order to improve neural network ’ s generalization ability and dynamic knowledge acquirement adaptive ability, a structure auto - adaptive neural network new model based on genetic algorithm is proposed to optimize structure parameter of nn including hidden layer nodes, training epochs, initial weights, and so on ; secondly, through establishing integrating neural network and introducing data fusion technique, the integrality and precision of acquired knowledge is greatly improved. then aiming at the incompleteness and uncertainty problem consisting in the process of knowledge acquirement, knowledge acquirement method based on rough sets is explored to fulfill the rule extraction for intelligent diagnosis expert system, by completing missing value data and eliminating unnecessary attributes, discretization of continuous attribute, reducing redundancy, extracting rules in this thesis. finally, rough sets theory and neural network are combined to form rnn ( rough neural network ) model for acquiring knowledge, in which rough sets theory is employed to carry out some preprocessing and neural network is acted as one role of dynamic knowledge acquirement, and rnn can improve the speed and quality of knowledge acquirement greatly
本文首先討論了影響神經網路的泛化能力的因素，提出了一種新的結構自適應神經網路學習演算法，在新方法中，採用了遺傳演算法對神經網路的結構參數（隱層節點數、訓練精度、初始權值）進行優化，大大提高了神經網路的泛化能力和知識動態獲取自適應能力;其次，構造集成神經網路，引入數據融合演算法，實現了基於集成神經網路的融合診斷，有效地提高了知識獲取的全面性、完善性及精度;然後，針對知識獲取過程中所存在的不確定性、不完備性等問題，探討了運用粗糙集理論的知識獲取方法，通過缺損數據補齊、連續數據的離散、沖突消除、冗餘信息約簡、知識規則抽取等一系列的演算法實現了智能診斷的知識規則獲取;最後，將粗糙集理論與神經網路相結合，研究了粗糙集-神經網路的知識獲取方法。 -
Some preceding rain factors were list, then stepwise regression algorithm was employed to select the obvious factors from the list as the input of the bp networks. and the trial - and - error method is employed to define the number of the hidden layer nodes
論文列出若干個前期降雨量因子，利用逐步回歸演算法從中挑選出影響因素大的作為網路的輸入，通過「試錯法」確定隱節點數。 -
The network consists of three layers : the input layer, the hidden layer, and the output layer
文中的bp網路模型都是由三層構成：輸入層、隱含層、輸出層。 -
The complex - valued weights between hidden and output layer are updated by solving linear system based on finding the complex - valued weights between input and hidden layer
當輸入層和隱層之間的權值計算出來后，就可以通過求解線性方程組得到隱層和輸出層之間的權值。 -
The network has four layers. input layer has 16 nodes. the first hidden layer has 17 nodes ; the second hidden layer has 10 nodes and one output node. 37 projects " data is used in training samples
網路共有四層，輸入層節點數為16個，隱含層一的節點數為17個，隱含層二的節點數10為個，輸出層節點1個。 -
A new way used to decide the neuron ' s number of the hidden layer is proposed based on the analysis and on the experiential way proposed by others
對神經網路隱含層的作用進行分析，在此基礎上，借鑒有關文獻提出的經驗公式，提出了確定隱含層節點數的新方法。 -
In this kind of networks, rough neurons locate in the hidden - layer and they consist of three parts which generated by two hyperplanes partition universe. the hyperplanes are obtained by support vector machines
該方法引入多個類似於支持向量機的子神經網路，並將網路中的隱含層單元設計成由多組粗糙神經元構成的網路單元。 -
Bp model can quite improve the accuracy of pricing result ; 2. need to confirm the number of the hidden layer of neuron rationally ; 3. should confirm population size rationally while optimizing ann ; 4
基本結論如下： （ 1 ） bp模型能夠提高定價結果的準確度；武漢理工大學碩士學位論文（ 2 ）建模需要合理確定隱含層神經元數目; （ 3 ）優化網路應恰當確定初始群體規模; （ 4 ）優化網路需要合理設計交叉運算元。 -
These are employed for constructing and configuring fuzzy neural network, where the number of neurons of hidden layer of network is equated to the number of rules and the initial weights of network are configured by above factors
首先利用粗糙集理論對樣本數據進行初步規則獲取，並計算規則的依賴度和條件覆蓋度，然後根據這些規則進行網路設計，其中，網路隱層節點的數目等於規則的數目，初始網路權重由規則的依賴度和條件覆蓋度確定，最後用遺傳演算法對模糊神經網路參數進行優化。 -
Based on analyzing the relationship between linear separability and a connected set in boolean space, the particular effect of a restraining neuron in extraction of rules from a bnn is discussed, and that effect is explained through a example called a mis problem in boolean space. in this paper, a pattern match learning algorithm of bnns is proposed. when a bnn has been trained by the algorithm, all the binary neurons of hidden layer belong to one or more ls series, if the logical meanings of those ls series are clear, the knowledge in the bnn can be dug out
另一個研究成果是在分析線性可分和樣本連通性關系的基礎上，以mis問題為例，討論了抑制神經元在二進神經網路規則提取中的獨特作用，提出了二進神經網路的模式匹配學習演算法，採用這種演算法對布爾空間的樣本集合進行學習，得到的二進神經網路隱層神經元都歸屬於一類或幾類線性可分結構系，只要這幾類線性可分結構系的邏輯意義是清晰的，就可以分析整個學習結果的知識內涵。
分享友人