scholarly journals Expert judgement-based tuning of the system reliability neural network

2013 ◽  
Vol 21 (1) ◽  
pp. 28-34
Author(s):  
A. Brandowski ◽  
Hoang Nguyen ◽  
Wojciech Frąckowiak
Keyword(s):  
Neural Network ◽  
Correlation Analysis ◽  
System Reliability ◽  
Expert Judgement ◽  
Data Correlation ◽  
Ahp Method ◽  
Output Data ◽  
Input And Output ◽  
The Neural Network ◽  
Technical Systems

ABSTRACT The neural network tuning procedure applied to reliability analyses of anthrop technical systems, based on judgements of experts - experienced operating practicians. Numerical and linguistic elicitation of the judgements, analyses of the network input and output data correlation and of the AHP method processing deviation are presented. Example of data elicitation and correlation analysis of a reliability arrangement of the seagoing ship propulsion system are included to the article.

REPRESENTATION FOR DISPERSION FORMULA OF OPTICAL FIBER USING HYBRID TECHNIQUE

2008 ◽  
Vol 19 (02) ◽  
pp. 205-213 ◽  
Author(s):  
AMR RADI
Keyword(s):  
Neural Network ◽  
Refractive Index ◽  
Optical Fiber ◽  
Hybrid Technique ◽  
Training Set ◽  
Fiber System ◽  
Output Data ◽  
Input And Output ◽  
Dispersion Formula ◽  
The Neural Network

Genetic Algorithm (GA) has been used to find the optimal neural network (NN) solution (i.e., hybrid technique) which represents dispersion formula of optical fiber. An efficient NN has been designed by GA to simulate the dynamics of the optical fiber system which is nonlinear. Without any knowledge about the system, we have used the input and output data to build a prediction model by NN. The neural network has been trained to produce a function that describes nonlinear system which studies the dependence of the refractive index of the fiber core on the wavelength and temperature. The trained NN model shows a good performance in matching the trained distributions. The NN is then used to predict refractive index that is not presented in the training set. The predicted refractive index had been matched to the experimental data effectively.

scholarly journals Influence of Data Scaling and Normalization on Overall Neural Network Performances in Photoacoustics

2021 ◽  
Author(s):  
Miroslava Ivko Jordovic Pavlovic ◽  
Katarina Djordjevic ◽  
Zarko Cojbasic ◽  
Slobodanka Galovic ◽  
Marica Popovic ◽  
...  
Keyword(s):  
Neural Network ◽  
Input Data ◽  
Photoacoustic Signal ◽  
Output Data ◽  
Network Training ◽  
Input And Output ◽  
The Neural Network ◽  
Data Scaling ◽  
Logarithmic Scaling ◽  
Network Prediction

Abstract In this paper, the influence of the input and output data scaling and normalization on the neural network overall performances is investigated aimed at inverse problem-solving in photoacoustics of semiconductors. The logarithmic scaling of the photoacoustic signal amplitudes as input data and numerical scaling of the sample thermal parameters as output data are presented as useful tools trying to reach maximal network precision. Max and min-max normalizations to the input data are presented to change their numerical values in the dataset to common scales, without distorting differences. It was demonstrated in theory that the largest network prediction error of all targeted parameters is obtained by a network with non-scaled output data. Also, it was found out that the best network prediction was achieved with min-max normalization of the input data and network predicted output data scale within the range of [110]. Network training and prediction performances analyzed with experimental input data show that the benefits and improvements of input and output scaling and normalization are not guaranteed but are strongly dependent on a specific problem to be solved.

Approaches in Correlation Analysis and Application on Oil Field Data

2016 ◽  
Vol 841 ◽  
pp. 77-82 ◽  
Author(s):  
David Vališ ◽  
Libor Žák
Keyword(s):  
Correlation Analysis ◽  
Field Data ◽  
Combustion Engine ◽  
Correlation Coefficients ◽  
Oil Field ◽  
Diesel Combustion ◽  
Spearman Correlation ◽  
Output Data ◽  
Input And Output

The paper deals with selected approaches which unite several correlation analysis principles. Field data very often has various forms and contents. The comparison of different approaches will help to determine more precisely which correlation analysis is better for assessing input and output data. In this paper we introduce several correlation principles which can help to select the most suitable correlation approach. We present a traditional correlation analysis and compare it with Pearson and Spearman correlation coefficients. The value of our article lies in comparing several different approaches of the correlation analysis in which the oil field data from diesel combustion engine are used

Critical Knock Diagnosis for Gasoline Engines Based on Neural Network with Wavelet Transform and Fuzzy Clustering

2012 ◽  
Vol 455-456 ◽  
pp. 1084-1089
Author(s):  
Jian Guo Yang ◽  
Yan Yan Wang ◽  
Bo Lin
Keyword(s):  
Neural Network ◽  
Wavelet Transform ◽  
Gasoline Engine ◽  
Engine Performance ◽  
Discrete Wavelet ◽  
New Approach ◽  
Gasoline Engines ◽  
Input And Output ◽  
The Neural Network ◽  
The Common

. It is difficult to detect critical knock for a gasoline engine by the common method of knock diagnosis. In this paper, a new approach is presented to detect critical knock for gasoline engines. Based on this approach knock diagnosis consists of four steps. Firstly, discrete wavelet transform (DWT) is chosen as a pre-processor for a neural network to extract knock characteristic signals; Secondly, four characteristic factors are selected and calculated from knock characteristic signals; Thirdly, degree of memberships of the characteristic factors are calculated as the input and output of the neural network; and finally a RBF(Radial Basis Function) neural network is chosen, trained and applied to detect critical knock. Knock experiments were performed on a gasoline engine, and the application of the presented approach was studied. The results show that the presented method is practicable and can be applied to control the ignition of a gasoline engine working under critical knock which is admitted as an improved state of engine performance.

Application of Wavelet Neural Network in Cortisol Solubility

2011 ◽  
Vol 396-398 ◽  
pp. 711-715
Author(s):  
Jian Xin Chen ◽  
Xu Na Shi ◽  
Shu Chun Pang ◽  
Mei Jing Zhang ◽  
Sheng Yu Li
Keyword(s):  
Neural Network ◽  
Wavelet Transforms ◽  
Learning Algorithm ◽  
Wavelet Neural Network ◽  
Training Data ◽  
Robust Learning ◽  
Input And Output ◽  
The Neural Network ◽  
Adaptively Adjusting ◽  
Flow Of Information

Wavelet neural network(WNN) was applied to predicate the cortisol solubility. The model consists of a multilayer feedforward hierarchical structure, and the flow of information is directed from the input to the output layer by using wavelet transforms to achieve faster convergence. By adaptively adjusting the number of training data involved during training, an adaptive robust learning algorithm is derived for improvement of the efficiency of the network. The neural network was trained and simulated cortisol solubility with different input and output parameters. Simulation results confirmed that this approach gave more accurate predictions solubility.

scholarly journals Trust and socio-economic indicators: model including religion factor

2018 ◽  
Keyword(s):  
Neural Network ◽  
Correlation Analysis ◽  
Interpersonal Trust ◽  
Economic Indicators ◽  
Gdp Per Capita ◽  
General Index ◽  
The Neural Network ◽  
Per Capita ◽  
The Relationship ◽  
Working Hypotheses

The article is concerned with the following issues: definitions, indicators of trust were reviewed; the working hypotheses of the research were formed; the choice of factors related to the trust indices was made; cluster analysis of the relationship between individual trust indices and economic indicators was carried out; a correlation analysis of the relationship between individual trust indices and socio-cultural indicators was conducted; a neural network for modeling the general index of trust based on a well-founded set of economic and socio-cultural indicators was developed. The hypothesis about the influence of socio-cultural factors on trust and out of which there was distinguished a relation to a specific religion. By means of correlation analysis and neural networks, it was shown that Protestantism and Catholicism are the most significant religions that affect the general index of interpersonal trust. However, atheism has a more significant impact. Following 198 observations, each of which represented the country for a given year in the period from 1995 to 2014, the neural network produced satisfactory results in forecasting the total trust index on the basis of the following factors: GDP per capita, GINI coefficient, atheism (percentage of population, support such an attitude to religion). The neural network recognized 89.9% of the data and 90% of the test set indicating that the network got adjusted and could be used for modeling. The scatter diagram for a 5% error indicates that most of the data is within the required value. But it should be noted, that the model overestimates trust in Ukraine at the end of the analyzed period. This gives grounds for the assumption that in Ukraine there are additional factors that negatively affect interpersonal trust.

INTERCONNECTION WIRE-LENGTH ESTIMATION IN VLSI STANDARD CELL PLACEMENT VIA NEURAL NETWORK

1998 ◽  
Vol 07 (04) ◽  
pp. 443-451 ◽  
Author(s):  
ASHUTOSH SAXENA ◽  
SUJU M. GEORGE ◽  
P. RAMBABU
Keyword(s):  
Neural Network ◽  
Standard Cell ◽  
Wire Length ◽  
Placement Problem ◽  
Network Characteristics ◽  
Length Estimation ◽  
Input And Output ◽  
The Neural Network ◽  
Cell Placement ◽  
Inaccurate Estimate

Neural Network is used as a tool for estimating interconnection wire-length in VLSI standard cell placement problem. Conventional methods for estimating the interconnection wire-length viz., Bounding Rectangle method, provide inaccurate estimate of the interconnection wire-length and does not depict the interconnection procedure in a layout and separates routing and placement tasks distinctly. The proposed mechanism utilizes the neural network characteristics in understanding the functional mapping between input and output, to estimate the interconnection wire-length. Experiments were performed for different number of cells with varying complexity of interconnections. In all the cases, the performance of the Neural Network is found to be superior to the results obtained using Bounding Rectangle procedure.

scholarly journals System Identification of Neural Signal Transmission Based on Backpropagation Neural Network

2020 ◽  
Vol 2020 ◽  
pp. 1-8
Author(s):  
Xiangyu Li ◽  
Chunhua Yuan ◽  
Bonan Shan
Keyword(s):  
Neural Network ◽  
Bp Neural Network ◽  
Signal Transmission ◽  
Neural Model ◽  
Neural Firing ◽  
Output Data ◽  
Neural Network Learning ◽  
Network Learning ◽  
Network Training ◽  
Input And Output

The identification method of backpropagation (BP) neural network is adopted to approximate the mapping relation between input and output of neurons based on neural firing trajectory in this paper. In advance, the input and output data of neural model is used for BP neural network learning, so that the identified BP neural network can present the transfer characteristics of the model, which makes the network precisely predict the firing trajectory of the neural model. In addition, the method is applied to identify electrophysiological experimental data of real neurons, so that the output of the identified BP neural network can not only accurately fit the neural firing trajectories of neurons participating in the network training but also predict the firing trajectories and spike moments of neurons which are not involved in the training process with high accuracy.

Using One-Dimensional Linear Interpolation Method to Check Over-Fitting in Neural Network with Multi-Dimensional Inputs

2014 ◽  
Vol 496-500 ◽  
pp. 2228-2232 ◽  
Author(s):  
Bao Lien Hung ◽  
Hsin An Hung
Keyword(s):  
Neural Network ◽  
Input Data ◽  
Interpolation Method ◽  
Linear Interpolation ◽  
One Dimensional ◽  
Input And Output ◽  
The Neural Network ◽  
Neural Network Structure ◽  
Linear Interpolation Method ◽  
Fitting In

Neural networks are widely used to learn and predict the correlation between input and output. However, in the process of learning, the excessive reduction of errors may influence the validity of prediction, this phenomenon is called over-fitting. In order to address this problem, this study sequenced the input data into one-dimensional data for the neural network structure of multi-dimensional inputs, and used visual graphics to observe whether there is over-fitting. This method is called one-dimensional linear interpolation method. The result of example validation proved that the proposed method can provide specific graphical information effectively, and determine whether there is over-fitting.

Sign in / Sign up

Export Citation Format

Share Document