Structure and weight optimization of neural network based on CPA-MLR and its application in naphtha dry point soft sensor

Wang, Yang; Chen, Chao; Yan, Xuefeng
May 2013
Neural Computing & Applications;May2013 Supplement, Vol. 22, p75
Academic Journal
Structure and weight of neural networks play an important role in the predicting performance of neural networks. In order to overcome the main flaws of neural networks, such as under-fitting, over-fitting or wasting computational resource, correlation pruning algorithm combined with multiple linear regression (CPA-MLR) is proposed to optimize the structure and weight of neural networks. Firstly, an initial three-layer network with the maximum nodes of hidden layer is selected, and BP is employed to train it. Secondly, correlation analysis of the hidden-layer output is carried out to confirm the redundant hidden nodes. Thirdly, the redundant nodes will be deleted one by one, and a multiple linear regression model between the output of the hidden layer and the expected input of the output layer, which can be obtained through the inverse function of the output-layer node, is employed to obtain their optimal weight. Finally, the optimal structure of the neural networks, which is corresponding to the best predicting performance of the neural networks, is obtained. Further, a practical example, that is developing naphtha dry point soft sensor, is employed to illustrate the performance of CPA-MLR. The results show that the predicting performance of the soft sensor is improved and then decreased with deleting the redundant nodes, and the optimal predicting performance is obtained with the optimal hidden nodes.


Related Articles

  • STATISTICAL MODELING OF THE COPPER LOSSES IN THE REVERBERATORY FURNACE SLAG. �OR�EVIC, Predrag; �IVKOVIC, �ivan; MIHAJLOVIC, Ivan; �TRBAC, Nada // Metalurgia International;Oct2011, Vol. 16 Issue 10, p120 

    This article presents the results of the statistical modeling of the slag composition (SiO2, FeO, Fe3O4, CaO, Al2O3) and the content of copper in the matte based on the resulting copper content in the slag during smelting of the sulfide concentrates in the reverberatory furnace. Multiple linear...

  • Neural-Net-Based Predictive Modeling of Spout Eye Size in Steelmaking. Palaneeswaran, Ekambaram; Brooks, Geoffrey; Xu, Xiaodong // Metallurgical & Materials Transactions. Part B;Jun2012, Vol. 43 Issue 3, p571 

    Gas stirring is commonly used in pyrometallurgical vessels to enhance mass and heat transfer and to promote impurity removal. In the case of secondary steelmaking, the spout eye area is caused by the escape of the gas from the top of the smelt where the liquid steel is directly exposed to the...

  • Brief Study of Identification System Using Regression Neural Network Base on Delay Tap Structures. Sahroni, Alvin // GSTF Journal on Computing;Dec2013, Vol. 3 Issue 3, p7 

    this investigation aims to provide an overview and investigations of non-linear identification system using neural network approach. Nowadays, a lot of neural network approach was done to provide a satisfying identification system. Backpropagation scheme as the mainstream approach for developing...

  • Mining of Optimized Multi Relational Relation Patterns for Prediction System. Kumar, D. Vimal; Tamilarasi, A. // International Review on Computers & Software;Jun2013, Vol. 8 Issue 6, p1356 

    Prediction is arguably considered the main goal of data mining, with the greatest potential payoff. The two principal prediction problems are classification and regression. In many data mining tools that support classification tasks, training data are stored in a single table containing both the...

  • Constrained optimum surface roughness prediction in turning of X20Cr13 by coupling novel modified harmony search-based neural network and modified harmony search algorithm. Farshbaf Zinati, Reza; Razfar, Mohammad // International Journal of Advanced Manufacturing Technology;Jan2012, Vol. 58 Issue 1-4, p93 

    Nowadays, manufacturers rely on trustworthy methods to predict the optimal cutting conditions which result in the best surface roughness with respect to the fact that some constraining functions should not exceed their critical values because of current restrictions considering competition found...

  • Network Intrusion Detection Based on a General Regression Neural Network Optimized by an Improved Artificial Immune Algorithm. Wu, Jianfa; Peng, Dahao; Li, Zhuping; Zhao, Li; Ling, Huanzhang // PLoS ONE;Mar2015, Vol. 10 Issue 3, p1 

    To effectively and accurately detect and classify network intrusion data, this paper introduces a general regression neural network (GRNN) based on the artificial immune algorithm with elitist strategies (AIAE). The elitist archive and elitist crossover were combined with the artificial immune...

  • Combined Forecast of Calibration Interval Based on Linear Trend Model and Neural Network. CHEN Zhenlin; ZHAO Fang; ZHANG Xiao // Applied Mechanics & Materials;2014, Issue 635-637, p662 

    For realizing the dynamic optimization of measuring instrument calibration interval, predicting the history calibration data by modeling. First the improved moving average method is used to modeling and to predict the development trend of parameters. On the basis of this, BP network is used to...

  • Forecasting e-Learning Efficiency by Using Artificial Neural Networks and a Balanced Score Card. Halachev, Petar // World Academy of Science, Engineering & Technology;Nov2010, Issue 47, p563 

    No abstract available.

  • Neural Network.  // Encyclopedic Reference of Molecular Pharmacology;2004, p626 

    An encyclopedia entry for "neural network" is presented. It refers to a computer algorithm which is used in solving non-linear optimisation problems. It was derives by analogy to the way in which the information is processed by the interconnected, parallel structure of the brain.


Read the Article


Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics