Application of Various Optimization Methods for Calculating the Neural Network Error for the Diagnosis of Neuromuscular Diseases

2019 ◽  
Vol 25 (1) ◽  
pp. 41-45
Author(s):  
K. Sh. Ismayilova ◽  
Keyword(s):  
Neural Network ◽  
Neuromuscular Diseases ◽  
Optimization Methods ◽  
The Neural Network

Use of Neural Network and Discrete Wavelet Transformations in Estimation of Road Profile

Volume 2 ◽  
2004 ◽  
Author(s):  
Mohammad Durali ◽  
Alireza Kasaaizadeh
Keyword(s):  
Neural Network ◽  
Optimization Methods ◽  
Discrete Wavelet ◽  
Feed Forward Neural Network ◽  
The Road ◽  
The Neural Network ◽  
Road Profile ◽  
Road Roughness ◽  
Coding Method ◽  
Dynamic Experiments

This paper presents a method for estimation of road profile for automotive research applications with more accuracy and higher speed. Dynamic response of a car equipped with position and velocity sensors and driving on a sample road is used as basic data. A feed-forward neural network, trained with outputs from a car model in ADAMS, is used as the car inverse model. The neural network is capable of estimating the road roughness from the car response during test drives. The ADAMS model is corrected and validated using a series of dynamic experiments on the car, performed on a hydro-pulse test rig. The only problem in this approach, like other identification and optimization methods, is the large volume of generated data in time domain, acquired from car response during road test. This problem is solved using wavelet methods to code the acquired data. Unlike all frequency methods that eliminate a large portion of the data details during processing, the wavelet coding method restores most of the details, while the volume of the stored data is kept to a minimum. The results show that this method can estimate the road profile accurately and with great savings in processing time and effort.

Improving the Accuracy of Parts Manufactured by Single Point Incremental Forming

2013 ◽  
Vol 332 ◽  
pp. 443-448 ◽  
Author(s):  
Crina Radu ◽  
Ion Cristea ◽  
Eugen Herghelegiu ◽  
Stefan Tabacu
Keyword(s):  
Neural Network ◽  
Stainless Steel ◽  
Incremental Forming ◽  
Single Point ◽  
Optimization Methods ◽  
Single Point Incremental Forming ◽  
Surface Method ◽  
Good Ability ◽  
The Neural Network ◽  
Network Method

The aim of this paper is to enrich the knowledge related to the single point incremental forming (SPIF) process by evaluating the efficiency of two optimization methods - the response surface method and the neural network method - to improve the accuracy of manufactured parts by prescribing a proper combination of the process parameters. The analysis is performed for a double frustum of pyramid made by stainless steel. It was found a good ability of prediction of both methods, demonstrating their suitability for physical implementation in solving problems associated to the SPIF process.

A Neural Network-Based Prediction of the Weld Lines in Resin Transfer Molded Parts

1999 ◽  
Author(s):  
Faezeh Soltani ◽  
Souran Manoochehri
Keyword(s):  
Neural Network ◽  
Back Propagation ◽  
Simulation Models ◽  
Optimization Methods ◽  
Computational Time ◽  
Back Propagation Algorithm ◽  
Rtm Process ◽  
The Neural Network ◽  
Molded Parts ◽  
The Neural Networks

Abstract A model is developed to predict the weld lines in Resin Transfer Molding (RTM) process. In this model, the preforms are assumed to be thin flat with isotropic and orthotropic permeabilities. The position of the weld lines formed by multiple specified inlet ports are predicted using a neural network-based back propagation algorithm. The neural network was trained with data obtained from simulation and actual molding experimentation. Part geometry is decomposed into smaller sections based on the position of the weld lines. The variety of preforms and processing conditions are used to verify the model. Applying the neural networks reduced the amount of computational time by several orders of magnitude compared with simulations. The models developed in this study can be effectively utilized in iterative optimization methods where use of numerical simulation models is cumbersome.

Evolutionary Optimization and Use of Neural Network for Optimum Stamping Process Design for Minimum Springback

2002 ◽  
Vol 2 (1) ◽  
pp. 38-44 ◽  
Author(s):  
K. M. Liew ◽  
Tapabrata Ray ◽  
H. Tan , and ◽  
M. J. Tan
Keyword(s):  
Neural Network ◽  
Evolutionary Algorithm ◽  
Sheet Metal ◽  
Process Parameters ◽  
Minimization Problem ◽  
Sheet Metal Forming ◽  
Metal Forming ◽  
Optimization Problems ◽  
Optimization Methods ◽  
The Neural Network

Sheet metal forming is characterized by various process parameters such as the forming sequence, shapes of products and dies, friction parameters, forming speed etc. A designer is faced with the challenge of identifying optimal process parameters for minimum springback. Currently, a vast majority of such applications in practice are guided by trial and error and user experience. In this paper, we present two useful designer aids; an evolutionary algorithm and a neural network integrated evolutionary algorithm. We have taken a simple springback minimization problem to illustrate the methodology although the evolutionary algorithm is generic and capable of handling both single and multiobjective, unconstrained and constrained optimization problems. The springback minimization problem has been modeled as a discrete variable, unconstrained, single objective optimization problem and solved using both optimization methods. Both the algorithms are capable of generating multiple optimal solutions in a single run unlike most available optimization methods that provide a single solution. The neural network integrated evolutionary algorithm reduces the computational time significantly as the neural network approximates the springback instead of performing an actual springback computation. The results clearly indicate that both the algorithms are useful optimization tools that can be used to solve a variety of parametric optimization problems in the domain of sheet metal forming.

scholarly journals Application of neural networks to the analysis of consumer opinions

2019 ◽  
Vol 13 ◽  
pp. 310-314
Author(s):  
Roman Mysan ◽  
Ivan Loichuk ◽  
Małgorzata Plechawska-Wójcik
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Network Optimization ◽  
Optimization Methods ◽  
User Feedback ◽  
Text Data ◽  
Neural Network Optimization ◽  
The Neural Network ◽  
Test Application

This paper presents an analysis of the possibilities of using neural networks to classify text data in the form of comments. Moreover, results of research of two neural network optimization methods: Adam and Gradient are presented. The aim of the work is to conduct research on the behavior of the neural network depending on the change of parameters and the amount of data used to teach the neural network. To achieve the goal, a test application was created. It uses a neural network to display the overall assessment of the accommodation facility based on the added user feedback.

scholarly journals A Deep Ultraviolet Mode-locked Laser Based on a Neural Network

2020 ◽  
Vol 10 (1) ◽  
Author(s):  
Haoyuan Lu ◽  
Hao Xu ◽  
Jianye Zhao ◽  
Dong Hou
Keyword(s):  
Neural Network ◽  
Laser System ◽  
Frequency Doubling ◽  
Optimization Methods ◽  
Mode Locking ◽  
Multiple Parameters ◽  
The Neural Network ◽  
Deep Ultraviolet ◽  
Mode Locked Laser ◽  
Focusing Lens

AbstractDeep ultraviolet lasers based on the phenomenon of mode-locking have been used widely in many areas in recent years, for example, in semiconductors, the environment and biomedicine. In the development of a mode-locked deep ultraviolet laser, one of the most important aspects is to optimize the multiple parameters of the complex system. Traditional optimization methods require experimenters with more optimization experience, which limits the wide application of the lasers. In this study, we optimize the deep ultraviolet mode-locked laser system using an online neural network to solve this problem. The neural network helps us control the position of the crystal, the length of the cavity, the position of the focusing lens and the temperature of the frequency doubling crystal. We generate a deep ultraviolet mode-locked laser with a power of 18 mW and a spectral center at 205 nm. This result is greatly improved compared to previous results with the same pump power. This technology provides a universal solution to multiparameter problems in the optimization of lasers.

scholarly journals Accelerating the learning process of a neural network by predicting the weight coefficient

2021 ◽  
Vol 4 (4) ◽  
pp. 295-302
Author(s):  
Viktor O. Speranskyy ◽  
Mihail O. Domanciuc
Keyword(s):  
Neural Network ◽  
Learning Process ◽  
Prediction Models ◽  
Optimization Methods ◽  
Weight Coefficient ◽  
Optimization Models ◽  
Initial Training ◽  
The Neural Network ◽  
Combined Use ◽  
Weight Coefficients

The purpose of this study is to analyze and implement the acceleration of the neural network learning process by predicting the weight coefficients. The relevance of accelerating the learning of neural networks is touched upon, as well as the possibility of using prediction models in a wide range of tasks where it is necessary to build fast classifiers. When data is received from the array of sensors of a chemical unit in real time, it is necessary to be able to predict changes and change the operating parameters. After assessment, this should be done as quickly as possible in order to promptly change the current structure and state of the resulting substances.. Work on speeding up classifiers usually focuses on speeding up the applied classifier. The calculation of the predicted values of the weight coefficients is carried out using the calculation of the value using the known prediction models. The possibility of the combined use of prediction models and optimization models was tested to accelerate the learning process of a neural network. The scientific novelty of the study lies in the effectiveness analysis of prediction models use in training neural networks. For the experimental evaluation of the effectiveness of prediction models use, the classification problem was chosen. To solve the experimental problem, the type of neural network “multilayer perceptron” was chosen. The experiment is divided into several stages: initial training of the neural network without a model, and then using prediction models; initial training of a neural network without an optimization method, and then using optimization methods; initial training of the neural network using combinations of prediction models and optimization methods; measuring the relative error of using prediction models, optimization methods and combined use. Models such as “Seasonal Linear Regression”, “Simple Moving Average”, and “Jump” were used in the experiment. The “Jump” model was proposed and developed based on the results of observing the dependence of changes in the values of the weighting coefficient on the epoch. Methods such as “Adagrad”, “Adadelta”, “Adam” were chosen for training neural and subsequent verification of the combined use of prediction models with optimization methods. As a result of the study, the effectiveness of the use of prediction models in predicting the weight coefficients of a neural network has been revealed. The idea is proposed and models are used that can significantly reduce the training time of a neural network. The idea of using prediction models is that the model of the change in the weight coefficient from the epoch is a time series, which in turn tends to a certain value. As a result of the study, it was found that it is possible to combine prediction models and optimization models. Also, prediction models do not interfere with optimization models, since they do not affect the formula of the training itself, as a result of which it is possible to achieve rapid training of the neural network. In the practical part of the work, two known prediction models and the proposed developed model were used. As a result of the experiment, operating conditions were determined using prediction models.

Neural Network for Automatic Analysis of Motility Data

1994 ◽  
Vol 33 (01) ◽  
pp. 157-160 ◽  
Author(s):  
S. Kruse-Andersen ◽  
J. Kolberg ◽  
E. Jakobsen
Keyword(s):  
Neural Network ◽  
Muscular Activity ◽  
Recording System ◽  
Automatic Analysis ◽  
Biologically Relevant ◽  
System A ◽  
The Neural Network ◽  
Valuable Method ◽  
Motor Abnormalities ◽  
Trained Network

Abstract:Continuous recording of intraluminal pressures for extended periods of time is currently regarded as a valuable method for detection of esophageal motor abnormalities. A subsequent automatic analysis of the resulting motility data relies on strict mathematical criteria for recognition of pressure events. Due to great variation in events, this method often fails to detect biologically relevant pressure variations. We have tried to develop a new concept for recognition of pressure events based on a neural network. Pressures were recorded for over 23 hours in 29 normal volunteers by means of a portable data recording system. A number of pressure events and non-events were selected from 9 recordings and used for training the network. The performance of the trained network was then verified on recordings from the remaining 20 volunteers. The accuracy and sensitivity of the two systems were comparable. However, the neural network recognized pressure peaks clearly generated by muscular activity that had escaped detection by the conventional program. In conclusion, we believe that neu-rocomputing has potential advantages for automatic analysis of gastrointestinal motility data.

Decision Support for Psychiatric Diagnosis Based on a Simple Questionnaire

1997 ◽  
Vol 36 (04/05) ◽  
pp. 349-351
Author(s):  
H. Mizuta ◽  
K. Kawachi ◽  
H. Yoshida ◽  
K. Iida ◽  
Y. Okubo ◽  
...  
Keyword(s):  
Neural Network ◽  
Decision Support ◽  
Psychiatric Diagnosis ◽  
Psychiatric Patients ◽  
Bayesian Classifier ◽  
Neural Network Classifier ◽  
Correct Decision ◽  
Neurotic Disorders ◽  
The Neural Network

Abstract:This paper compares two classifiers: Pseudo Bayesian and Neural Network for assisting in making diagnoses of psychiatric patients based on a simple yes/no questionnaire which is provided at the outpatient’s first visit to the hospital. The classifiers categorize patients into three most commonly seen ICD classes, i.e. schizophrenic, emotional and neurotic disorders. One hundred completed questionnaires were utilized for constructing and evaluating the classifiers. Average correct decision rates were 73.3% for the Pseudo Bayesian Classifier and 77.3% for the Neural Network classifier. These rates were higher than the rate which an experienced psychiatrist achieved based on the same restricted data as the classifiers utilized. These classifiers may be effectively utilized for assisting psychiatrists in making their final diagnoses.

Sign in / Sign up

Export Citation Format

Share Document