Multilayer Backpropagation Research Articles

Machine Learning Model for Dynamical Response of Nano-composite Pipe Conveying Fluid under Seismic Loading

Machine learning approaches including support vector regression (SVR) and multi-layer feedforward backpropagation neural network (FFBNN) were used in the present study along with classic theory for predicting maximum displacement of nanocomposite pipe conveying fluid under seismic load. The FFBNN consisted of three layers: 1) three neurons in input layer including length-to-radius ratio (L/R), fluid velocity (V) and volume percent of carbon nanotube; 2) hidden layer with 11 neurons obtained via trial and error; 3) maximum displacement-based seismic load. SVR model was obtained via three-input data with maximum likelihood estimator. Model predicted results were compared using three metrics, including Nash-Sutcliffe efficiency, root mean squared error and coefficient of correlation for 100 testing and 255 training data points. Results indicated that SVR achieved best predictions in the training phase, while FFBNN provided superior prediction in the testing phase. Increasing L/R, V and decreasing VCNT, increased maximum displacements under seismic load.

Determination of positioning accuracies by using fingerprint localisation and artificial neural networks

Fingerprint localisation technique is an effective positioning technique to determine the object locations by using radio signal strength, values in indoors. The technique is subject to big positioning errors due to challenging environmental conditions. In this paper, initially, a fingerprint localisation technique is deployed by using classical k-nearest neighborhood method to determine the unknown object locations. Additionally, several artificial neural networks, are employed, using fingerprint data, such as single-layer feed forward neural network, multi-layer feed forward neural network, multi-layer back propagation neural network, general regression neural network, and deep neural network to determine the same unknown object locations. Fingerprint database is built by received signal strength indicator measurement signatures across the grid locations. The construction and the adapted approach of different neural networks using the fingerprint data are described. The results of them are compared with the classical k-nearest neighborhood method and it was found that deep neural network was the best neural network technique providing the maximum positioning accuracies.

A Multilayered Back Propagation Algorithm to Predict Significant Attributes of UG Pursuing Students Absenteeism at Rural Educational Institution

International Journal of Computer Sciences and Engineering (A UGC Approved and indexed with DOI, ICI and Approved, DPI Digital Library) is one of the leading and growing open access, peer-reviewed, monthly, and scientific research journal for scientists, engineers, research scholars, and academicians, which gains a foothold in Asia and opens to the world, aims to publish original, theoretical and practical advances in Computer Science,Information Technology, Engineering (Software, Mechanical, Civil, Electronics & Electrical), and all interdisciplinary streams of Computing Sciences. It intends to disseminate original, scientific, theoretical or applied research in the field of Computer Sciences and allied fields. It provides a platform for publishing results and research with a strong empirical component. It aims to bridge the significant gap between research and practice by promoting the publication of original, novel, industry-relevant research.

Backpropagation to predict drinking water consumption

The present research article is based on the use of artificial neural networks as a tool for the prediction applied to the consumption of drinking water, where the artificial learning of a multilayer backpropagation network with the historical data consumed in m3 is used. In a computer with a regular feature, the backpropagation network was implemented in the Python 3.7.0 language, taking as a study object a case published on its SUNEDU website regarding drinking water consumption in 2017 and checking the results in January of 2018, taking these data for learning and the respective tests. It has been achieved as a result to predict the amount of water consumption of the institution. The test performed resulted in an excess of 23 m3, which represents 2.7% in excess given that in January 2018 a consumption of 833m3 was recorded. Bear in mind that it was done for the training process with a maximum error for the training of 0.000099 and with a maximum number of iterations of 100000. After training with the historical data of drinking water consumption, it was possible to predict the consumption of SUNEDU to January 2018.

Automated recognition and classification of adulteration levels from bulk paddy grain samples

Fraudulent labeling and adulteration are the major concerns in the global rice industry. Almost all the paddy varieties being sold in the market are prone to adulteration. It is very difficult to differentiate paddy grains of various varieties in the mixed bulk sample based on visual observation. Currently, there is no sophisticated appearance-based commercial scale technology to reliably detect and quantify adulteration in bulk paddy grain samples. The paper presents a cost-effective image processing technique for the recognition of adulteration and classification of adulteration levels (%) from the images of adulterated bulk paddy samples using state-of-the-art color and texture features. In this work, seven adulterated bulk paddy samples are considered and each of the samples is prepared by mixing a premium paddy variety with the identical looking and commercially inferior paddy variety at five different adulteration levels (weight ratios) of 10%, 15%, 20%, 25% and 30%. The study compares the performances of three different classification models, namely, Multilayer Back Propagation Neural Network (BPNN), Support Vector Machine (SVM) and k-Nearest Neighbor (k-NN). The Principal Component Analysis (PCA) and Sequential Forward Floating Selection (SFFS) methods have been employed separately for the automatic selection of optimal feature subsets from the combined color and texture features. The maximum average adulteration level classification accuracy of 93.31% is obtained using the BPNN classification model trained with PCA-based reduced features. The proposed technique can be used as an economic, rapid, non-destructive and quantitative technique for testing adulteration, authenticity, and quality of bulk paddy grain samples.

Artificial neural networks model for estimating growth of polyculture microalgae in an open raceway pond

Microalgae have potential as biomass energy sources with higher photosynthetic efficiency compared to terrestrial plants. The use of polyculture systems such as native microalgae communities for microalgae cultivation has several advantages, as well as challenges due to indeterminate species composition and growth rate variation between species. This paper presents an artificial neural network (ANN) model to estimate the growth of polyculture microalgae in a semi-continuous open raceway pond (ORP). The model was comprised of a multilayer backpropagation neural network with eight input parameters, one hidden layer, and one output parameter. The model was developed using datasets collected from the cultivation of polyculture microalgae in Minamisoma City, Fukushima Prefecture, Japan. The input parameters are as follows: initial algal concentration, harvesting period (between two and three days after the growth have begun), hydraulic retention time, addition of sodium acetate, average solar radiation (μmole m−2 s−1), average temperature (oC), pH condition, and nitrate ion ( NO 3 − ) concentration. The output variable is the microalgae concentration observed during the cultivation period. The output is represented using a single neuron. The result of the study showed that the designed three-layer ANN achieved a high prediction accuracy (R2 = 0.93) for all combinations of inputs.

Effects of Climate Change on Soil Erosion Risk Assessed by Clustering and Artificial Neural Network

The erosivity index, as a combination of the Fournier index (FI) and the Bagnouls–Gaussen aridity index (BGI), has been suggested to assess soil erosion risk. It can be easily calculated from meteorological data, i.e., precipitation and temperature. As an example application, data from 55 meteorological stations in Turkey corresponding to a period of 39 years from 1975 to 2013 are considered herein. The stations were classified using cluster analysis to obtain a zonation of Turkey based on EI yearly averages. Clustering techniques were applied to a similarity matrix between stations obtained based on the complement of the probability of the similarity ratio index. Four clusters were defined according to the maximal evenness of the eigenvalues of the within each cluster similarity matrices corresponding to different hierarchical levels of the dendrograms. The probability of similarity was calculated using a permutation technique. Time series of the EI of the clusters were used to predict their annual average values for the years from 2014 to 2040 using a multilayer back-propagation neural network (MLPBP-NN). The results showed that the four clusters represent a gradient of increasing EI. The clusters corresponding to northern and central Turkey have lower EI values and EI variability than those for southern and western Turkey. The results of the MLPBP-NN predict that the erosion risk will increase for all zones, but with high increments in southern and western Turkey. Therefore, the regions corresponding to these clusters should be subjected to detailed soil erosion risk analysis.

Early warning modeling and analysis based on a deep radial basis function neural network integrating an analytic hierarchy process: A case study for food safety

Food safety is vital to the national economy and livelihood of people. Therefore, effective food safety warnings are helpful to the healthy and sustainable development of society. Focused on the early warning modeling for a certain scale of complex food safety inspection data, this paper proposes a novel early warning modeling method based on the deep radial basis function (DRBF) neural network that integrates an analytic hierarchy process (AHP). First, the AHP based on the entropy weight is used to obtain the risk fusion results of the inspection data as the expected output of the DRBF. Then, the DRBF model based on the autoencoder is used to build the early warning model, implementing feature learning to acquire the high-level representation of the food inspection data. Finally, the category data of sterilized milk from the food safety inspection data of a province in China is taken as a case study. Comparing the experimental results of the radial basis function (RBF) neural network, the back propagation (BP) neural network and the improved multilayer BP, the proposed DRBF model is found to have a better generalization ability and a better generalization effect for the complex food safety inspection data. Furthermore, the proposed early warning model is used to predict and analyze the risk of the inspection data from early September 2014. The results could be helpful for relevant departments to carry out early warning work and provide a scientific basis for guidance, thereby promoting the improvement of the food quality and reducing food risks.

Online Learning Styles Identification Model, Based on the Analysis of User Interactions Within an E-Learning Platforms, Using Neural Networks and Fuzzy Logic

Individual Learning Style identification is an essential aspect in the development of intelligent or adaptive e-Learning platforms. Traditional methods are based on the application of questionnaires or psychological tests, which may not be the most appropriate in all cases. The proposed model is based on the analysis of user behavior through the study of their interactions within an e-Learning platform, using a multilayer Backpropagation Neural Network and Fuzzy Logic concepts, for the preprocessing of the inputs and the categorization of the outputs.

Assessment of Settlement Profile Caused by Underground Box Structure Installation with an Artificial Neural Network Model

Various empirical equations have been proposed to predict the ground settlement profile caused by the excavation of conventional circular tunnels. However, ground movement for the underground box structure has not been fully studied. In this study, ground settlement induced by underground box installations is investigated using two-dimensional finite element analyses. A new formulation to assess the settlement profile applicable to underground box structure is proposed based on parametric analyses of the changes in ground condition, geometric condition of structure, and construction conditions. This paper also presents a method to predict the maximum surface settlement around an underground box structure with artificial neural networks (ANNs), taking into account nine input variables that have direct physical significance. A MATLAB-based multi-layer back propagation neural network model is developed, trained, and tested with parameters obtained from numerical analyses. The maximum settlement from the ANN model, in conjunction with a new formulation to construct the settlement profile, turns out to be promising, by predicting a settlement profile compatible with field measurement data.

Wavelet transform based power quality events classification using artificial neural network and SVM

This paper demonstrates classification of PQ events utilizing wavelet transform (WT) energy features by artificial neural network (ANN) and SVM classifiers. The proposed scheme utilizes wavelet based feature extraction to be used for the artificial neural networks in the classification. Six different PQ events are considered in this study. Three types of neural network classifiers such as feed forward multilayer back propagation (FFML), learning vector quantization (LVQ), probabilistic neural network (PNN) are analyzed for effective classification of PQ events. The results show the superiority of PNN over FFML and LVQ. The test simulations show that SVM has higher performance than ANN with feed forward multilayer back propagation (FFML), learning vector quantization (LVQ), probabilistic neural network (PNN).Keywords: Power Quality, Daubechies, Multi Resolution Analysis, Feature Extraction, Neural networks, SVM

A multilayer backpropagation saliency detection algorithm and its applications

Saliency detection is an active topic in the multimedia field. Most previous works on saliency detection focus on 2D images. However, these methods are not robust against complex scenes which contain multiple objects or complex backgrounds. Recently, depth information supplies a powerful cue for saliency detection. In this paper, we propose a multilayer backpropagation saliency detection algorithm based on depth mining by which we exploit depth cue from three different layers of images. The proposed algorithm shows a good performance and maintains the robustness in complex situations. Experiments' results show that the proposed framework is superior to other existing saliency approaches. Besides, we give two innovative applications by this algorithm, such as scene reconstruction from multiple images and small target object detection in video.

A Seasonal Model Using Optimized Multi-Layer Neural Networks to Forecast Power Output of PV Plants

With the continuous increase of grid-connected photovoltaic (PV) installed capacity and the urgent demand of synergetic utilization with the other power generation forms, the high-precision prediction of PV power generation is increasingly important for the optimal scheduling and safe operation of the grid. In order to improve the power prediction accuracy, using the response characteristics of PV array under different environmental conditions, a data driven multi-model power prediction method for PV power generation is proposed, based on the seasonal meteorological features. Firstly, through the analysis of PV power characteristics in typical seasons and seasonal distribution of the weather factors, such as solar irradiance and ambient temperature, the influences of different weather factors on PV power prediction are studied. Then, according to the meteorology characteristics of Beijing, different seasons can be divided. The historical data corresponding to different seasons are acquired and then the seasonal PV power forecasting models are established based on optimized multi-layer back propagation neural network (BPNN), realizing the multi-model prediction of PV power. Finally, effectiveness of the seasonal PV power forecasting method is compared and validated. The performance analysis of the neural network forecasting model under typical seasonal conditions shows that the multi-model forecasting method based on seasonal characteristics of PV power generation is better than that of single power forecasting model for the whole year. The results show that the proposed method can effectively improve the power forecasting accuracy of PV power.

Stacked Sparse Auto-Encoders (SSAE) Based Electronic Nose for Chinese Liquors Classification.

This paper presents a stacked sparse auto-encoder (SSAE) based deep learning method for an electronic nose (e-nose) system to classify different brands of Chinese liquors. It is well known that preprocessing; feature extraction (generation and reduction) are necessary steps in traditional data-processing methods for e-noses. However, these steps are complicated and empirical because there is no uniform rule for choosing appropriate methods from many different options. The main advantage of SSAE is that it can automatically learn features from the original sensor data without the steps of preprocessing and feature extraction; which can greatly simplify data processing procedures for e-noses. To identify different brands of Chinese liquors; an SSAE based multi-layer back propagation neural network (BPNN) is constructed. Seven kinds of strong-flavor Chinese liquors were selected for a self-designed e-nose to test the performance of the proposed method. Experimental results show that the proposed method outperforms the traditional methods.

Diagnosing Parkinson's Disease Using Movement Signal Mapping by Neural Network and Classifier Modulation

Parkinson's disease is a growing and chronic movement disorder, and its diagnosis is difficult especially at the initial stages. In this paper, movement characteristics extracted by a computer using multilayer back propagation neural network mapping are converted to the symptoms of this disease. Then, modulation of three classifiers of C4.5, k-nearest neighbors, and support vector machine with majority voting are applied to support experts in diagnosing the disease. The purpose of this study is to choose appropriate characteristics and increase the accuracy of the diagnosis. Experiments were performed to demonstrate the improvement of Parkinson's disease diagnosis using this method.

A novel conjugate gradient method with generalized Armijo search for efficient training of feedforward neural networks

Abstract In this paper, a novel multilayer backpropagation (BP) neural network model is proposed based on conjugate gradient (CG) method with generalized Armijo search. The presented algorithm requires low memory and performs fast convergent speed in practical applications. One reason is that the constructed conjugate direction guarantees the sufficient descent behavior in minimizing the given objective function. The other stems from the fact that the generalized Armijo method can automatically determine a more suitable learning rate in each training epoch. As a theoretical contribution, two deterministic convergent results, weak and strong convergence, have been detailedly proved under more relaxed assumptions. The weak convergence means that the norm of gradient of the objective function tends to zero. For the strong convergence, it represents that the sequence of weight vectors approaches a fixed point. To support the theoretical results, some illustrated simulations have been done on various benchmark datasets.

Parametric study of closed wet cooling tower thermal performance

The present study involves experimental and theoretical analysis to evaluate the thermal performance of modified Closed Wet Cooling Tower (CWCT). The experimental study includes: design, manufacture and testing prototype of a modified counter flow forced draft CWCT. The modification based on addition packing to the conventional CWCT. A series of experiments was carried out at different operational parameters. In view of energy analysis, the thermal performance parameters of the tower are: cooling range, tower approach, cooling capacity, thermal efficiency, heat and mass transfer coefficients. The theoretical study included develops Artificial Neural Network (ANN) models to predicting various thermal performance parameters of the tower. Utilizing experimental data for training and testing, the models simulated by multi-layer back propagation algorithm for varying all operational parameters stated in experimental test.

Prediction of axial capacity of piles driven in non-cohesive soils based on neural networks approach

This paper presents an application of two advanced approaches, Artificial Neural Networks (ANN) and Princi­pal Component Analysis (PCA) in predicting the axial pile capacity. The combination of these two approaches allowed the development of an ANN model that provides more accurate axial capacity predictions. The model makes use of Back-Propagation Multi-Layer Perceptron (BPMLP) with Bayesian Regularization (BR), and it is established through the incorporation of approximately 415 data sets obtained from data published in the literature for a wide range of un-cemented soils and pile configurations. The compiled database includes, respectively 247 and 168 loading tests on large-and low-displacement driven piles. The contributions of the soil above and below pile toe to the pile base resistance are pre-evaluated using separate finite element (FE) analyses. The assessment of the predictive performance of the new method against a number of traditional SPT-based approaches indicates that the developed model has attractive capabili­ties and advantages that render it a promising tool. To facilitate its use, the developed model is translated into simple design equations based on statistical approaches.

Optimization of Machining Parameters for Minimizing Surface Roughness in Turning GFRP Composite Using ANN and PSO Methodology

The influence of surface roughness in determining the quality of finished products in any industrial application has an enormous impact on gaining competitive edge and establishing superiority. Thus, recognizing and understanding the factors influencing the resulted surface roughness are the crucial issues helping to achieve the desired goal in any competitive industrial environment. Fact is that machining process parameters are major factors affecting the outcome. This research is focused on determining the optimum machining parameters (cutting speed, feed rate, depth of cut) which result in minimizing the surface roughness in turning glass fiber reinforced polymer (GFRP) matrix composite using coated carbide insert. To understand the effects of machining parameters on surface roughness and to determine relationship between them; Particle Swarm Optimization (PSO) has been employed. A multiple regression equation is used as objective function to determine the optimum values of inputs (cutting speed, feed, and depth of cut) using PSO formula and it yields an optimum value of surface roughness of 0.6252 µm. Artificial Neural Network (ANN) has also been implemented to predict various level of surface roughness for different machining parameters. To predict the surface roughness (Ra), standard multilayer feed-forward back-propagation hierarchical neural network has been applied and the findings provide an overall value of coefficient of determination of 0.88881. These investigations of turning operation provide optimal process parameters for any desired value of surface roughness which result in gaining a competitive edge over others in any industrial application.

Optimized Neural Network for Spectrum Prediction Scheme in Cognitive Radio

Cognitive radio technologies permit the sharing of spectrums between unlicensed as well as licensed customers, based on the principle of non-interference. Spectrum sensing, in this field, is hence a vital tool that assists in the ascertainment of availability of a particular channel within the licensed spectrum, for unlicensed customers. But this function uses significant power that could be lessened through the employment of predictive mechanisms to discover spectrum holes. The traffic features of licensed customer systems in the real world are not known beforehand. In this paper, a spectrum predictor based on Neural Networks model Multi-Layer Perceptron and Back Propagation that do not need prior information regarding traffic features of licensed customers is designed. Binary Shuffled Frog Leaping Algorithm is proposed for structure optimization, the binary structure is suggested to show the memes with the purpose of developing a sub-collection with lesser dimensions than that of the original collection where detecting sensitivity and accuracy would be scalable with that of the primary status. Spectrum Predictor's performance is examined through exhaustive experiments.