Forward Neural Network Research Articles

Application of Back Propagation Neural Network Model for Predicting Flank Wear of Yttria Based Zirconia Toughened Alumina (ZTA) Ceramic Inserts

A back propagation neural network model has been adopted for the flank wear prediction of zirconia toughened alumina (ZTA) insert in turning operation. The experiments are performed on AISI 4340 steel using developed yttria based ZTA inserts. These inserts are prepared through wet chemical co-precipitation route followed by powder metallurgy process. Machining conditions such as cutting speed, feed rate and depth of cut are selected as input to the neural network model and flank wear of the inserts corresponding to these conditions has been chosen as the output of the network. The experimentally measured values are used to train the feed forward back propagation artificial neural network for prediction of those conditions. The convergence of the mean square error both in training and testing come out very well. The performance of the trained neural network has been validated with experimental data. The results demonstrate that the machining model is suitable and the optimization strategy satisfies practical requirements.

Modeling of Surface Roughness in Turning Operation Using Extreme Learning Machine

Prediction model allows the machinist to determine the values of the cutting performance before machining. According to the literature, various modeling techniques have been investigated and applied to predict the cutting parameters. Response surface methodology (RSM) is a statistical method that only predicts effectively within the observed data provided. Most artificial intelligent systems mostly had an issue with user-defined data and long processing time. Recently, the extreme learning machine (ELM) method has been introduced, combining the single hidden layer feed- forward neural network with analytically determined output weights. The advantage of this method is that it can overcome the limitations due to the previous methods which include too many engineers’ judgment and slow iterative learning phase. Therefore, in this study, the ELM was proposed to model the surface roughness based on RSM design of experiment. The results indicate that ELM can yield satisfactory solution for predicting the response within a few seconds and with small amount of error.

The Application of Multi Layer Feed Forward Artificial Neural Network for Learning Style Identification

Accommodating learning styles in adaptive educational hypermedia systems (AEHS) improves students' learning performance in web-based learning. Hence, in implementing the systems, the students' learning style identification process is important. Most of today's AEHS rely on a traditional technique of identifying students' learning styles, which is using questionnaires. However, using questionnaires for this purpose is less efficient, cumbersome and may not be that feasible. This study proposes a real-time learning style identification technique by recording students' browsing behaviors and analyzing them by using multi-layer feed forward artificial neural network (MLFF). The result suggests that there is a relationship between the frequencies of students' click on learning components with their staying time on those components. It also indicates that the proposed identification technique performs well in identifying students' learning styles.

Effect of Geometrical Factors on Magnetic Induction Distribution of Toroidal Cores Using Numerical Methods

Toroidal cores wound from grain-oriented 3 % SiFe 0.27-mm thick (M4) electrical steel are widely used in many applications. Apart from the processing factors, the geometry of the toroid has an effect on its magnetic properties. In magnetic path length of a toroid from its inside to its outside circumference, the magnetic field varies at different layers. Artificial neural network have been successfully used for the prediction of magnetic performance in electromagnetic devices. Experimental data obtained from the previous measurements have been used as a training data to a feed forward multilayer perceptron neural network for the prediction of magnetic field and flux density distribution. The input parameters were outer and inner diameters and strip thickness while the output parameter was the magnetic field and flux density. When the network was tested by untrained sample data, the average correlation of the models was found to be 99 % and the overall prediction error was in the range of 2.31 and 0.02 in Ampere per meter and Tesla. An analytical equation as depending on theoretical data and prediction results has been determined by using MATLABⓇ Curve Fitting Toolbox TM for magnetic field and flux density distribution in the toroid.

The Analysis of BP Network Traffic Accident Serious Level Model Based on Principal Component

This paper analyzes the relevant factor which influences the accident severity, and puts forward BP neural network traffic accident serious level model based on principal component analysis, it selects the main causes and accident serious level of 344 traffic accidents in Harbin in 2003 on the main road as the training and binding of the model, make sure that there is a good correspondence between the related main causes and the serious level. The model provides reference for the main control factors in order to determine the serious level of road traffic accidents of Nonlinear Systems.

Evolved neural network ensemble by multiple heterogeneous swarm intelligence

The neural network ensemble (NNE) is a very effective way to obtain a good prediction performance by combining the outputs of several independently trained neural networks. Swarm intelligence is applied here to model the population of interacting agents or swarms that are able to self-organize. In this paper, we combine NNE and multi-population swarm intelligence to construct our improved neural network ensemble (INNE). First, each component forward neural network (FNN) is optimized by chaotic particle swarm optimization (CPSO) and gradient gescending (GD) algorithm. Second, in contrast to most existing NNE training algorithm, we adopt multiple obviously different populations to construct swarm intelligence. As an example, one population is trained by particle swarm optimization (PSO) and the others are trained by differential evolution (DE) or artificial bee colony algorithm (ABC). The ensemble weights are trained by multi-population co-evolution PSO-ABC-DE chaotic searching algorithm (M-PSO-ABC-DE-CS). Our experiments demonstrate that the proposed novel INNE algorithm is superior to existing popular NNE in function prediction.

Neural Network-Based Prediction of Effective Heat Storage Coefficient of Building Materials

In the present paper, we have employed the application of artificial neural networks (ANN) to predict effective heat storage coefficient (HSC) of building materials. First we prepared a database to train and test the models developed here. Two types of architectures from different networks are developed, one with three inputs and the other with four inputs mixed architecture combining an ANN with a theoretical model developed by us previously. These ANN models are built, trained and tested by the feed forward back propagation algorithm, to obtain the effective properties of building materials from the properties of their constituents. Feed forward back propagation neural network structure has been developed, which includes an input layer, a hidden layer and an output layer. The number of neurons in the input layer is equal to the number of input parameters and the number of neurons in the output layer is equal to the output parameters. A good agreement has been found between the predicted values using ANN and the experimental results reported in the literature.

Outlier detection based on the neural network for tensor estimation

Rationale and objectivesDiffusion weighted imaging (DWI) is always influenced by both thermal noise and spatially/temporally varying artifacts such as subject motion and cardiac pulsation. Motion artifacts are particularly prevalent, especially when scanning an uncooperative population with several disorders. Some motion between acquisitions can be corrected by co-registration approaches. However, automated and accurate motion outlier detection of brain DWIs is an integral component of the analysis and interpretation of tensor estimation. Many different and innovative methods have been proposed to improve upon this technology. In this study, we proposed a classifier work frame, which can classify DWIs as normal images or motion artifacts. Materials and methodsThe procedure contains the following stages: first, we used the wavelet transform to extract features from the original DWIs; second, the principle component analysis was used to reduce the features; third, the forward neural network (FNN) was employed to construct the classifier; fourth, a Rossler-based chaotic particle swarm optimization method was proposed to train the FNN; fifth, the cost matrix was determined as the false negative (FN) which was 10 times larger than the false position (FP); and finally, the K-fold cross validation was chosen to avoid overfitting. We applied this method on 60 DWI datasets, including 50 training datasets and 10 test datasets. ResultsThe experimental results based on our DWI database showed that the proposed method can effectively extract the global feature from images and achieve better performance in tensor estimation by automatic unvoxelwise outlier rejection compared with manual and visual inspection, and previous voxelwise outlier rejection methods. We found that the motion artifact detection accuracy on both the training and test datasets was over 95.8%, while the computation time per DWI slice was only 0.0149s. ConclusionThe proposed method could potentially remove the influence of unexpected motion artifacts in DWI acquisitions and should be applicable to other magnetic resonance imaging.

An Approach of an Image Watermarking Scheme using Neural Network

of an image watermarking scheme using neural network is presented in this work. In the proposed approach, watermark comes from the weights of an identify image that are loaded from a learned feed-forward neural network, the neural network is learned by using the back-propagation learning algorithm. To improve the robustness of watermarked image; the procedure of watermark embedding is embedded into host image through selecting and modifying the Gaussian coefficients comes from a noisy image. The noised image is damaged by the salt and pepper noise. In order to identify the cover of extracted watermark, feed- forward neural network is used in the watermarking identification to overcome the limitation of unknown data comes randomly. The results of the scheme realization show the robustness of proposed scheme that has preferable performance for both identification and watermarking of a noised image

Investigation of ANFIS and FFBNN Recognition Methods Performance in Tamil Speech Word Recognition

In speech word recognition systems, feature extraction and recognition plays a most significant role. More number of feature extraction and recognition methods are available in the existing speech word recognition systems. In most recent Tamil speech word recognition system has given high speech word recognition performance with PAC-ANFIS compared to the earlier Tamil speech word recognition systems. So the investigation of speech word recognition by various recognition methods is needed to prove their performance in the speech word recognition. This paper presents the investigation process with well known Artificial Intelligence method as Feed Forward Back Propagation Neural Network (FFBNN) and Adaptive Neuro Fuzzy Inference System (ANFIS). The Tamil speech word recognition system with PAC-FFBNN performance is analyzed in terms of statistical measures and Word Recognition Rate (WRR) and compared with PAC-ANFIS and other existing Tamil speech word recognition systems.

Specific Gravity, RVP, Octane Number, and Saturates, Olefins, and Aromatics Fractional Composition of Gasoline and Petroleum Fractions by Neural Network Algorithms

An accurate method for estimation of the properties of petroleum fractions based on their ASTM D86 distillation profile is developed using a set of conventional feed forward single layer neural networks algorithms. Results on 362 gasoline and petroleum fuels show that the proposed method outperforms other alternatives in terms of accuracy, simplicity, and generality with overall correlation coefficients 0.995 for research octane number, 0.994 for motor octane number, 0.995 for Reid vapor pressure, 0.9996 for specific gravity, and 0.994 for saturates 0.992 for olefins and 0.998 for aromatics fractional compositions. The model is suitable for incorporation into ASTM D86 test apparatus to predict the properties of petroleum fractions in a single test, resulting in savings in terms of space, time, energy, and cost.

Rapid and Precise GLONASS GDOP Approximation using Neural Networks

Russian global navigation satellite system (GLONASS) provides civilian and military users three-dimensional position determination and navigation services as same as US global positioning system. Geometric dilution of precision (GDOP) provides a simple interpretation of positioning precision. Usual method for GLONASS GDOP calculation is matrix inversion. However this process imposes a huge calculation load on receiver, especially when large number of visible satellites exists. To overcome this problem, artificial neural network is used. Different configurations and training methods are simulated on a data base obtained by a GLONASS receiver. Then navigation precision and execution times are explored and compared. Results show that recurrent neural network has 0.00024 RMS error, which is the best against other focused tools including feed forward back propagation and radial basis function neural network with usual training and with genetic algorithm adopted weights and biases.

Artificial Neural Network trained by Genetic Algorithm for Smart MIMO Channel Estimation for Downlink LTE-Advance System

Long-Term Evolution (LTE) is the next generation of current mobile telecommunication networks. LTE has a flat radio-network architecture and significant increase in spectrum efficiency, throughput and user capacity. In this paper, performance analysis of robust channel estimators for Downlink Long Term Evolution-Advanced (DL LTE-A) system using three Artificial Neural Networks: Feed-forward neural network (FFNN), Cascade-forward neural network (CFNN) and Layered Recurrent Neural Network (LRN) are trained separately using Back-Propagation Algorithm and also ANN is trained by Genetic Algorithm (GA). The methods use the information got by the received reference symbols to estimate the total frequency response of the channel in two important phases. In the first phase, the proposed ANN based method learns to adapt to the channel variations, and in the second phase it estimates the channel matrix to improve performance of LTE. The performance of the estimation methods is evaluated by simulations in Vienna LTE-A DL Link Level Simulator in MATLAB software. Performance of the proposed channel estimator, ANN trained by Genetic Algorithm (ANN- GA) is compared with traditional Least Square (LS) algorithm and ANN based other estimator like Feed- forward neural network, Layered Recurrent Neural Network and Cascade-forward neural network for Closed Loop Spatial Multiplexing (CLSM)-Single User Multi-input Multi-output (MIMO-2×2 and 4×4) in terms of throughput. Simulation result shows proposed ANN- GA gives better performance than other ANN based estimations methods and LS. Index Terms — LTE-A, MIMO, Artificial Neural Network (ANN), Back-Propagation, Layered Recurrent Neural Network (LRN), Feed-forward neural network (FFNN), Cascade-forward neural network (CFNN), Genetic Algorithm (GA).

OFDM Link with a Better Performance Using Artificial Neural Network

In this paper, a scheme for reducing the peak-to-average power ratio (PAPR) primarily and bit error rate simultaneously in orthogonal frequency division multiplexing using feed forward artificial neural network is proposed and analyzed. The proposed scheme switches one or more null-carriers with data subcarriers to offer minimum PAPR and requires no channel side information during transmission. To reduce high computational complexity, a multilayer neural network with Levenberg–Marquardt training algorithm is used. The proposed scheme thus suits to supplement some of other PAPR-reduction methods with low rate hit and complexity.

Prediction of resonant frequency of a circular patch frequency selective structure using artificial neural network

In this paper, efforts have been given to develop a method which relates resonant frequency with the periodicity of a frequency selective structure comprising of two dimensional arrays of circular patches. Firstly resonant frequencies for various periodicities of the frequency selective surfaces have been calculated by method of moment after selecting proper basis function. Results obtained by this method have been used to train a feed forward back propagation artificial neural network with three hidden layers. Genetic algorithm based back propagation neural network and gradient decent based artificial neural networks are used in this process. From the trained network, resonant frequencies for given periodicities (even excluding the training data set) may be readily available. Now for the same combination of periodicity resonant frequency has been again calculated by method of moment. Some of these results are compared with the measured data for validation. Each time it is found that the results are in good parity.

Classification of power quality disturbances using minimum entropy and feed forward back propagation neural network approach

Classification of power quality disturbances using minimum entropy and feed forward back propagation neural network approach

Nonlinear Ethanol Gasoline Optimal Control System Based on Hammerstein Model

In this paper, a Hammerstein model based on forward feedback neural networks was proposed to tackle the optimal control of a nonlinear MISO system. The method offers a solution to the optimization of internal models. The optimal control with the preset value was implemented under both static and dynamic optimal indices. The simulation results showed that the algorithm can fulfill the task of blending ethanol and gasoline effectively.

Compatible abnormality detection technique for CT and MRI brain images

An automated computerised tomography (CT) and magnetic resonance imaging (MRI) brain images are used to perform an efficient classification. The proposed technique consists of three stages, namely, pre-processing, feature extraction and classification. Initially, pre-processing is performed to remove the noise from the medical MRI images. Then, in the feature extraction stage, the features that are related with MRI and CT images are extracted and these extracted features which are given to the Feed Forward Back-propagation Neural Network (FFBNN) is exploited in order to classify the brain MRI and CT images into two types: normal and abnormal. The FFBNN is well trained by the extracted features and uses the unknown medical brain MRI images for classification in order to achieve better classification performance. The proposed method is validated by various MRI and CT scan images. A classification with an accomplishment of 96% and 70% has been obtained by the proposed FFBNN classifier. This achievement shows the effectiveness of the proposed brain image classification technique when compared with other recent research works.

Neural Network Modeling of Methanol to Propylene over P-ZSM-5 in a Fluidized Bed

P-ZSM-5 catalyst selectively converts methanol to propylene (MTP) and other olefins at moderate temperatures (325– 550˚C). Between 250–325˚C, the catalyst dehydrates methanol to dimethyl ether (MTD) reaching 87% at low gas velocities in a fluidized bed. A carbon pool forms in the catalyst pores from which propylene and other hydrocarbons are derived. Propylene selectivity varied between 20–41% at temperatures between 400–550˚C and weight hourly space velocities (WHSV) of 0.20–0.39 h −1 . Two multilayer feed forward artificial neural network (ANN) models represent the methanol conversion and propylene selectivity very well ( 22 0.973 and 0.999 MTD MTP RR = = , respectively). Temperature is the most significant factor: conver sion increases from 42% at 300 ˚C to 63% at about 500˚C. The rank of importance for conversion and selectivity are: catalyst weight >argon flow >methanol flow and water flow >catalyst weight >methanol flow >argon flow, respectively. ANN is an alternative to kinetic and hydrodynamic modeling to optimize reactor performance particularly for complex systems such as MTP in fluidized beds.

SPECTRAL DOMAIN FEATURES FOR OVARIAN CANCER DATA ANALYSIS

The early detection of cancer is crucial for succes sful treatment. Medical researchers have investigat ed a number of early-diagnosis techniques. Recently, the y have discovered that some cancers affect the concentration of certain molecules in the blood, wh ich allows early diagnosis by analyzing the blood m ass spectrum. Researchers have developed several techniques for the analysis of the mass-spectrum curve analysis and used them for the detection of prostat e, ovarian, breast, bladder, pancreatic, kidney, li ver and colon cancers. In this study we propose a new techn ique that uses the spectral domain features such as wavelet transform and Fourier transform for the ana lysis of the ovarian cancer data to differentiate b etween normal and patients with malignant cancer. We used two different classifiers for the original data, th e first one is a feed forward artificial neural network cla ssifier which gave a sensitivity of 96%, specificit y of 88% and accuracy of 94%. The second used classifier is the linear discriminant analysis classifier which separated the cancer from healthy samples with sens itivity of 79%, specificity of 75% and accuracy of about 81%. After transforming the data to the spectral do main using the Fourier transform the performance wa s degraded. The experimental results showed that the performance of the wavelet transform based system was superior to other techniques as it gave a sensitivi ty of 98%, specificity of 96% and accuracy of 95%.