uln

Compared with traditional neural network , the structure of ULN is more compact ;

同传统的神经网络相比，通用学习网络的结构更加紧凑；

The identification with ULN establishes the base for effective control of pH neutralization process .

通用学习网络对pH中和过程的辨识为对该过程的有效控制奠定了基础。

In this paper , Universal Learning Network ( ULN ), is used to identify the typical nonlinear and long time delay system .

本文将通用学习网络（UniversalLearningNetwork）应用于对非线性、大滞后系统的辨识。

It was important for the theoretical and practical value of the occurrence and control of ULN in lilies .

该研究对百合叶烧机理的发生和防治具有重要的理论和实际应用价值。

And how the generalization ability of modeling dynamic systems is influenced by the network size is also studied using the ULN with switching mechanism .

还用带有开关机制的ULN对网络规模是如何影响动态系统泛化能力的问题进行了研究。

A new modeling method using the ULN with switching mechanism is studied , where both parameters and time delays are adjusted to model the nonlinear systems .

研究了一种新的带有开关机制的ULN建模方法，不仅对其中的参数而且对时间延迟都进行了调整以适应非线性系统的建模要求。

Has the proof of the liver harm , the ALT > 1.5 × ULN .

有肝损伤的证据，ALT>1.5×ULN；

Methods The cases were chronic hepatitis-B , with positive HBV-DNA , positive HBeAg and ALT ≤ 1 ~ 2 ULN ( upper limit of normal ) .

方法病例为慢性乙型肝炎伴有乙型肝炎病毒（HBV）DNA阳性、乙型肝炎病毒e抗原（HBeAg）阳性及ALT≤1～2正常上限值（ULN）。

When ULN is adopted in the system with long time-delay , the network can embody the long time-delay properly , which establish the base for ULN to identity nonlinear dynamical system with long time-delay .

应用于大滞后系统时，系统的延迟时间能够在网络中得到充分体现，为通用学习网络有效地辨识大滞后动态系统奠定了基础。

The research on this problem is developed in the thesis using artificial neural network theory , especially Universal Learning Network ( ULN ) used in this field the first time , and the method of identification and control based on the ULN is presented .

本文利用人工神经网络对这一问题进行了研究，特别是将通用学习网络（UniversalLearningNetwork，ULN）首次应用于这一领域，初步提出了基于通用学习网络的辨识和控制方法。

A universal learning network ( ULN ), of which the nodes are connected with one another and there are multiple branches with arbitrary time delays between any two nodes , is used to identify the typical nonlinear and long time delay system .

利用通用学习网络具有所有节点互连，任意两节点之间可以有多重连接，且连接允许有任意延迟时间的特点，对典型非线性、大滞后系统进行了辨识。

A new type of neural network & universal learning network ( ULN ) is introduced , the nodes of the network is connected each other , and between any two nodes , there are multi-branch , and also the time-delay can be set arbitrarily on each branch .

通用学习网络（UniversalLearningNetwork）是一种新型可设定神经网络，该网络所有节点互连，任意两节点之间可以有多重连接，且连接上有任意的延迟时间。

A new type of neural network - universal learning network ( ULN ) is used to identify this process . The results show that both the training error and the generalizing error have been reduced , and the output of the trained network fits the sample signal very well .

利用一种新型的神经网络&通用学习网络对该过程进行辨识，取得了良好的辨识结果，网络训练误差和泛化误差均达到较高精度，网络训练输出曲线与教师信号曲线拟合良好。

Results Those ALT > 3 × ULN , HBV DNA < 4.2 × 10 ~ 6 copies / ml before the treatment , and with virus response at an early stage have a high HBeAg seroconversion ( P < 0.01 or < 0.05 ) .

结果治疗前ALT≥3×ULN，HBVdna<4.2×106拷贝/ml，出现早期病毒学应答者HBeAg血清转换率较高（P<0.01或<0.05）。