Artificial Neural Network Learning Algorithms Research Articles

Replacement of numerical simulations with machine learning in the inverse problem of two-phase flow in porous medium

A possibility to replace full-physics numerical simulations with machine-learning-based algorithms in inverse problems of multiphase flow in porous media is studied on an example of specialized oil-well tests. The performance function is computed at each iteration of the inverse problem solution using an artificial neural network (ANN) instead of forward numerical simulations. Proper ANN structure, training set, and learning algorithm are established and implemented. Achieved approximation quality for the performance function on a test sample is well suitable for the inverse problem solution. Further improvements and alternative ANN formulations are proposed for practical applications.

The influence of algorithms for tuning the parameters of neuromorphic systems on their fault tolerance

This article describes the influence of algorithms for tuning the parameters of neuromorphic systems on their fault tolerance. This is relevant to the hardware implementation of neuromorphic systems using memristors (NSM). The study is conducted using the authors developed a variant of the system approach and methods of simulation of artificial neural networks (ANN). By the example of a multilayer perceptron, it is shown that different ANN learning algorithms in the nominal mode of operation make it possible to achieve similar values of the operation accuracy. But due to the influence of production and operational factors in real conditions of operation, the ANN may fail. The range of allowable values of the destabilizing factors on ANN operation depends on the learning algorithm and may differ several times.

Rainfall-Runoff Model Based on ANN with LM, BR and PSO as Learning Algorithms

Rainfall-runoff model requires comprehensive computation as its relation is a complex natural phenomenon. Various inter-related processes are involved with factors such as rainfall intensity, geomorphology, climatic and landscape are all affecting runoff response. In general there is no single rainfall-runoff model that can cater to all flood prediction system with varying topological area. Hence, there is a vital need to have custom-tailored prediction model with specific range of data, type of perimeter and antecedent hour of prediction to meet the necessity of the locality. In an attempt to model a reliable rainfall-runoff system for a flood-prone area in Malaysia, 3 different approach of Artificial Neural Networks (ANN) are modelled based on the data acquired from Sungai Pahang, Pekan. In this paper, the ANN rainfall-runoff models are trained by the Levenberg Marquardt (LM), Bayesian Regularization (BR) and Particle Swarm Optimization (PSO). The performances of the learning algorithms are compared and evaluated based on a 12-hour prediction model. The results demonstrate that LM produces the best model. It outperforms BR and PSO in terms of convergence rate, lowest mean square error (MSE) and optimum coefficeint of correlation. Furthermore, the LM approach are free from overfitting, which is a crucial concern in conventional ANN learning algorithm. Our case study takes the data of rainfall and runoff from the year 2012 to 2014. This is a case study in Pahang river basin, Pekan, Malaysia.

A novel approach for efficient crop yield prediction

Crop yield prediction is one of the challenging task in agricultural domain. Extensive research in agricultural domain has been carried out to predict better crop yield using the machine learning algorithm Artificial Neural Network (ANN) and statistical model Multiple Linear Regression (MLR). This article examines the intrinsic relationship between MLR and ANN. A hybrid MLR-ANN model has been proposed in this research work for efficient crop yield prediction. The proposed hybrid model is modeled to analyze the prediction accuracy when MLR intercept and coefficients were applied to initialize the ANN’s input layer weights and bias. Feed Forward Artificial Neural Network with Back Propagation training algorithm was used for predicting accurate paddy crop yield. In conventional ANN model, the weights and bias of input and hidden layer are initilized randomly. This hybrid MLR-ANN model, instead of random weights and bias initialization, the input layer weights and bias are initialized by using MLR’s coefficients and bias. The hybrid model prediction accuracy is compared with ANN, MLR, Support Vector Regression (SVR), k-Nearest Neighbour (KNN) and Random Forest (RF) models by using performance metrics. The computational time for both hybrid MLR-ANN and conventional ANN was calculated. The results show that the proposed hybrid MLR-ANN model gives better accuracy than the conventional models.

Quantitative CT texture analysis for diagnosing systemic sclerosis: Effect of iterative reconstructions and radiation doses.

To test whether texture analysis (TA) can discriminate between Systemic Sclerosis (SSc) and non-SSc patients in computed tomography (CT) with different radiation doses and reconstruction algorithms.In this IRB-approved retrospective study, 85 CT scans at different radiation doses [49 standard dose CT (SDCT) with a volume CT dose index (CTDIvol) of 4.86 ± 2.1 mGy and 36 low-dose (LDCT) with a CTDIvol of 2.5 ± 1.5 mGy] were selected; 61 patients had Ssc ("cases"), and 24 patients had no SSc ("controls"). CT scans were reconstructed with filtered-back projection (FBP) and with sinogram-affirmed iterative reconstruction (SAFIRE) algorithms. 304 TA features were extracted from each manually drawn region-of-interest at 6 pre-defined levels: at the midpoint between lung apices and tracheal carina, at the level of the tracheal carina, and 4 between the carina and pleural recesses. Each TA feature was averaged between these 6 pre-defined levels and was used as input in the machine learning algorithm artificial neural network (ANN) with backpropagation (MultilayerPerceptron) for differentiating between SSc and non-SSc patients.Results were compared regarding correctly/incorrectly classified instances and ROC-AUCs.ANN correctly classified individuals in 93.8% (AUC = 0.981) of FBP-LDCT, in 78.5% (AUC = 0.859) of FBP-SDCT, in 91.1% (AUC = 0.922) of SAFIRE3-LDCT and 75.7% (AUC = 0.815) of SAFIRE3-SDCT, in 88.1% (AUC = 0.929) of SAFIRE5-LDCT and 74% (AUC = 0.815) of SAFIRE5-SDCT.Quantitative TA-based discrimination of CT of SSc patients is possible showing highest discriminatory power in FBP-LDCT images.

Machine Learning Prediction Models for In-Hospital Mortality After Transcatheter Aortic Valve Replacement

ObjectivesThis study sought to develop and compare an array of machine learning methods to predict in-hospital mortality after transcatheter aortic valve replacement (TAVR) in the United States. BackgroundExisting risk prediction tools for in-hospital complications in patients undergoing TAVR have been designed using statistical modeling approaches and have certain limitations. MethodsPatient data were obtained from the National Inpatient Sample database from 2012 to 2015. The data were randomly divided into a development cohort (n = 7,615) and a validation cohort (n = 3,268). Logistic regression, artificial neural network, naive Bayes, and random forest machine learning algorithms were applied to obtain in-hospital mortality prediction models. ResultsA total of 10,883 TAVRs were analyzed in our study. The overall in-hospital mortality was 3.6%. Overall, prediction models’ performance measured by area under the curve were good (>0.80). The best model was obtained by logistic regression (area under the curve: 0.92; 95% confidence interval: 0.89 to 0.95). Most obtained models plateaued after introducing 10 variables. Acute kidney injury was the main predictor of in-hospital mortality ranked with the highest mean importance in all the models. The National Inpatient Sample TAVR score showed the best discrimination among available TAVR prediction scores. ConclusionsMachine learning methods can generate robust models to predict in-hospital mortality for TAVR. The National Inpatient Sample TAVR score should be considered for prognosis and shared decision making in TAVR patients.

A neuro-particle swarm optimization logistic model fitting algorithm for software reliability analysis

This article develops a particle swarm optimization algorithm based on a feed-forward neural network architecture to fit software reliability growth models. We employ adaptive inertia weight within the proposed particle swarm optimization in consideration of learning algorithm. The dynamic adaptive nature of proposed prior best particle swarm optimization prevents the algorithm from becoming trapped in local optima. These neuro-prior best particle swarm optimization algorithms were applied to a popular flexible logistic growth curve as the [Formula: see text] model based on the weights derived by the artificial neural network learning algorithm. We propose the prior best particle swarm optimization algorithm to train the network for application to three different software failure data sets. The new search strategy improves the rate of convergence because it retains information on the prior particle, thereby enabling better predictions. The results are verified through testing approaching of constant, modified, and linear inertia weight. We assess the fitness of each particle according to the normalized root mean squared error which updates the best particle and velocity to accelerate convergence to an optimal solution. Experimental results demonstrate that the proposed [Formula: see text] model based prior best Particle Swarm Optimization based on Neural Network (pPSONN) improves predictive quality over the [Formula: see text], [Formula: see text], and existing model.

Development of a Driving Behavior-Based Collision Warning System Using a Neural Network

An advanced driver assistance system (ADAS) uses radar, visual information, and laser sensors to calculate variables representing driving conditions, such as time-to-collision (TTC) and time headway (THW), and to determine collision risk using empirically set thresholds. However, the empirically set threshold can generate differences in performance that are detected by the driver. It is appropriate to quickly relay collision risk to drivers whose response speed to dangerous situations is relatively slow and who drive defensively. However, for drivers whose response speed is relatively fast and who drive actively, it may be better not to provide a warning if they are aware of the collision risk in advance, because giving collision warnings too frequently can lower the reliability of the warnings and cause dissatisfaction in the driver, or promote disregard. To solve this problem, this study proposes a collision warning system (CWS) based on an individual driver’s driving behavior. In particular, a driver behavior model was created using an artificial neural network learning algorithm so that the collision risk could be determined according to the driving characteristics of the driver. Finally, the driver behavior model was learned using actual vehicle driving data and the applicability of the proposed CWS was verified through simulation.

Neuro-Fuzzy Hybridization using Modified S Membership Function and Kernel Extreme Learning Machine for Robust Face Recognition under Varying Illuminations

The multifaceted light varying environment severely degrades the performance of person recognition using facial images. Here, the authors present a novel person identification method using hybridization of artificial neural network (ANN) and fuzzy logic concepts. An efficient illumination normalization method is presented with the help of a new modified S membership function. The proposed method of illumination normalization retains the large scale facial features as well as suppresses the variations related to change in light variations. Kernel extreme learning machine (KELM) which is a nonlinear and non-iterative learning algorithm of ANN is used for classification. Various kernel types and parameters are experimented to find the best choice for robust classification. To assess the performance of proposed hybridization, Yale and extended Yale B face databases have been used. Very promising results have been achieved which establish the worth of the proposed method.

Epileptic seizure anticipation and localisation of epileptogenic region using EEG signals

Electric activity of brain gets disturbed prior to epileptic seizure onset. Early prediction of an upcoming seizure can help to increase effectiveness of antiepileptic drugs. The scalp electroencephalogram signals contain information about the dynamics of brain and have been used to predict an upcoming seizure and localise its zone. The objective of this paper is to localise the epileptogenic region and predict an upcoming seizure at the earliest. To localise epileptogenic region, Electroencephalogram signals are categorised into four regions of brain (Frontal, Temporal, Parietal and Central). For each signal seventy-two (72) parameters in frequency domain have been extracted by using ten minute non overlapping window. Four prominent ratio parameters, γ1/γ5, γ3/γ1, θ/γ2 and γ4/θ have been identified as best parameters based on relative fisher score. Zone 2 shows the highest change in all the parameters as compared to the other zones. So, temporal region is identified as the epileptogenic region in this work. For prediction of the epileptic seizure machine learning algorithm artificial neural network (ANN) is proposed. The proposed machine learning algorithm has an accuracy of 92.3%, sensitivity of 100% and specificity of 83.3%.

Integrated Big Data Analytics Technique for Real-Time Prognostics, Fault Detection and Identification for Complex Systems

Real-time prediction of the state of complex systems is vital for integrity management since it is easier to plan for asset maintenance, reduce risks associated with unplanned downtime and reduce the cost of maintenance. This study utilized a four-fold cross-validation ensemble for an Artificial Neural Network (ANN) that used Multi-Layer Perceptron (MLP) in a backward propagation technique for haul crane prognosis. Big data on components’ degradation states obtained from the Supervisory Control And Data Acquisition (SCADA) systems were used to implement the study. After preprocessing the dataset, importance scoring was used to compute the Cumulative Target-component Percentage-influence (CTP) of the input variables (source components) on the output variable (the target component) at the 95.5%, 99.3%, 99.9% and 100% levels. The specific source components responsible for the CTP levels of the target component were later used for the ANN network training that followed the cross-validation ensemble technique. The cross-validation ensemble ANN technique was also compared to the classic ANN and other machining learning algorithms. Finally, the best-trained cross-validation ensemble ANN network, which was obtained at the 99.9% CTP level, was used for future estimation of the time of failure of the system to enhance planning for the expected maintenance program that will be required at such times.

A Novel Computer Assisted Automatic Sleep Scoring System by Employing Artificial Neural Network–A review

Sleep is an essential element for an individual’s well-being and is considered vital for the overall mental and physical heath of a person. Sleep can be considered as a virtual detachment of an individual from his environment. In normal humans, about 30% of their life-time is spent for sleep. Artificial neural networks (ANNs) or connectionist systems are computing systems inspired by the biological neural networks that constitute animal brains. Sleep scoring is under taken by the examination and visual inspection of polysomnograms (PSG) done by sleep specialist. PSG is specialty test, the conduction of which includes the recording of various physiological signals. The signals obtained are processed using digital processing tools so as to extract information. Soft computing techniques are used to analyze the signals. ANNs are parallel adaptive systems suitable for solving of non-linear problems. Using ANN for automatic sleep scoring is especially promising because of new ANN learning algorithms allowing faster classification without decreasing the performance. Both appropriate preparations of training data as well as selection of the ANN model make it possible to perform effective and correct recognizing of relevant sleep stages. Such an approach is highly topical, taking into consideration the fact that there is no automatic scorer utilizing ANN technology available at present. The high performances observed with systems based onneural networks highlight that these tools may be act new tools in the field of sleep research. In this scenario we are surmised the review regarding the computer assisted automatic scoring of sleep and soft computing technique Artificial Neural Network.

A two stroke free piston engine’s performance and exhaust emission using artificial neural networks

The performance and exhaust emissions of a Free piston linear engine (FPLE) were ascertained for various equivalence ratios (0.7, 0.8, 0.9, 1.0, 1.1 and 1.2). After that, the usability of Artificial neural networks (ANNs) in case of FPLE has been tested. Actually, the aim was to examine the best suited operational condition of, and show the possibility of using ANN, for this kind of engine technology. We first interrogated the thermal efficiency, generated power, total heat release rate, indicated mean effective pressure, exhaust gas temperature, and exhaust emissions such as CO, CO2, NOx and O2 for the chosen range of equivalence ratios and then gathered experimental data in order to train and test an artificial neural network model for prediction. The experimental results showed that, running the engine on the slightly lean side stoichiometrically, can fulfil the goals of higher engine performance and lower emissions. We used the back propagation learning algorithm for ANN and observed that the correlation coefficients (R) vary between 0.990-0.999, Mean absolute percentage error (MAPE) vary between 0.885-5.9%, and coefficient of determination (R2) vary between 0.937-0.999; which showed a welldefined relationship between the predicted and experimental values.

Comparison of linear regression and artificial neural network model of a diesel engine fueled with biodiesel-alcohol mixtures

This study deals with usage of linear regression (LR) and artificial neural network (ANN) modeling to predict engine performance; torque and exhaust emissions; and carbon monoxide, oxides of nitrogen (CO, NOx) of a naturally aspirated diesel engine fueled with standard diesel, peanut biodiesel (PME) and biodiesel-alcohol (EME, MME, PME) mixtures. Experimental work was conducted to obtain data to train and test the models. Backpropagation algorithm was used as a learning algorithm of ANN in the multilayered feedforward networks. Engine speed (rpm) and fuel properties, cetane number (CN), lower heating value (LHV) and density (ρ) were used as input parameters in order to predict performance and emission parameters. It was shown that while linear regression modeling approach was deficient to predict desired parameters, more accurate results were obtained with the usage of ANN.

TERMINATION CRITERION FOR PCA WITH ANN FOR DETECTION OF NS1 FROM ADULTERATED SALIVA

Detection of Non-structural Protein 1 (NS1) in saliva has become appealing as it may lead to a non-invasive detection method for NS1-related diseases at the febrile phase, before complication developed. NS1 is found to have its unique molecular fingerprint from Surface Enhanced Raman Spectroscopy (SERS) technique. Our work here intends to investigate the effect of termination criterion of Principal Component Analysis (PCA) on the classification performance by the different Artificial Neural Network (ANN) learning algorithms. This will help in optimizing the automated classification of NS1 adulterated saliva, and hence detection of NS1-related diseases. Raman spectra of saliva (n=64) and saliva mixed with NS1 (n=64) are acquired using SERS obtained from the UiTM-NMRR 12868-NS1-DENV database. Large input data dimension of the raw [128 x 1801] are reduced according to the respective PCA termination criteria: Scree test [128 x 5], Cumulative Percent of Total Variance (CPV) [128 x 70] and Eigenvalues One Criterion (EOC) [128 x 115]. The reduced data dimensions are used as inputs to ANN algorithms. Performance of these algorithms, in term of [accuracy, sensitivity, specificity, and precision] from Levenbergh Marquardt (LM), Scale Conjugate Gradient (SCG), Resilient Backpropagation (RPROP) and One Step Secant (OSS) are investigated. The best performance [100%, 100%, 100%, 100%] are achieved from the integration of Scree test criterion and SCG learning algorithm, the highest expected of adulterated data.

A Framework for Detecting Fraudulent Activities in EDO State Tax Collection System Using Investigative Data Mining

The Inland Revenue Services is overwhelmed with gigabyte of disk capacity containing data about tax payers in the state. The data stored on the database increases in size at an alarming rate. This has resulted in a data rich but information poor situation where there is a widening gap between the explosive growth of data and its types, and the ability to analyze and interpret it effectively, hence the need for a new generation of automated and intelligent tools and techniques known as investigative data mining, to look for patterns in data. These patterns can lead to new insights, competitive advantages for business, and tangible benefits for the State Revenue services. This research work focuses on designing effective fraud detection and deterring architecture using investigative data mining technique. The proposed system architecture is designed to reason using Artificial Neural Network and Machine learning algorithm in order to detect and deter fraudulent activities. We recommend that the architectural framework be developed using Object Oriented Programming and Agent Oriented Programming Languages.

Optimal low-thrust spiral trajectories using Lyapunov-based guidance

For an increasing number of electric propulsion systems used for real missions, it is very important to design optimal low-thrust spiral trajectories for these missions. However, it is particularly challenging to search for optimal low-thrust transfers. This paper describes an efficient optimal guidance scheme for the design of time-optimal and time-fixed fuel-optimal low-thrust spiral trajectories. The time-optimal solution is obtained with Lyapunov-based guidance, in which the artificial neural network (ANN) is adopted to implement control gains steering and the evolutionary algorithm is used as the learning algorithm for ANN. Moreover, the relative efficiency introduced in Q-law is analyzed and a periapis-and-apoapsis-centered burn structure is proposed for solving time-fixed fuel-optimal low-thrust orbit transfer problem. In this guidance scheme, the ANN is adopted to determine the burn structure within each orbital revolution and the optimal low-thrust orbit transfer problem is converted to the parameter optimization problem. This guidance scheme runs without an initial guess and provides closed form solutions. In addition, Earth J2 perturbation and Earth-shadow eclipse effects are considered in this paper. Finally, a comparison with solutions given by the literature demonstrates the effectiveness of the proposed method.

Artificial Neural Networks and Human Brain: Survey of Improvement Possibilities of Learning

<p class="R-AbstractKeywords">There are numerous applications of Artificial Neural Networks (ANN) at the present time and there are different learning algorithms, topologies, hybrid methods etc. It is strongly believed that ANN is built using human brain’s functioning principles but still ANN is very primitive and tricky way for real problem solving. In the recent years modern neurophysiology advanced to a big extent in understanding human brain functions and structure, however, there is a lack of this knowledge application to real ANN learning algorithms. Each learning algorithm and each network topology should be carefully developed to solve more or less complex problem in real life. One may say that almost each serious application requires its own network topology, algorithm and data pre-processing. This article presents a survey of several ways to improve ANN learning possibilities according to human brain structure and functioning, especially one example of this concept – neuroplasticity – automatic adaptation of ANN topology to problem domain.</p>

Comparative investigation of artificial neural network learning algorithms for modeling solar still production

Three artificial neural network learning algorithms were utilized to forecast the productivity (MD) of a solar still operating in a hyper-arid environment. The learning algorithms were the Levenberg–Marquardt (LM), the conjugate gradient backpropagation with Fletcher–Reeves restarts, and the resilient backpropagation. The Julian day, ambient air temperature, relative humidity, wind speed, solar radiation, temperature of feed water, temperature of brine water, total dissolved solids (TDS) of feed water, and TDS of brine water were used in the input layer of the developed neural network model. The MD was located in the output layer. The developed model for each algorithm was trained, tested, and validated with experimental data obtained from field experimental work. Findings revealed the developed model could be utilized to predict the MD with excellent accuracy. The LM algorithm (with a minimum root mean squared error and a maximum overall index of model performance) was found to be the best in the training, testing, and validation stages. Relative errors in the predicted MD values of the developed model using the LM algorithm were mostly in the vicinity of ±10%. These results indicated that the LM algorithm is the most ideal and accurate algorithm for the prediction of the MD with the developed model.

Adaptive behavior of structural systems in unpredictable changing environments by using self-learning algorithms: A case study

Adaptive architecture is expected to improve the performance of buildings and create more efficient building systems. One of the major research areas under this scope is the adaptive behavior of structural elements affected by load distribution. In order to achieve this, current studies develop structures that adapt by either following a database of pre-calculated equilibrium solutions or using self-learning algorithms to acquire active control systems to structures. This paper examined a case study element, which demonstrates an adaptive behavior in real time, based on self-learning abilities. The focus of this experiment was to gain control over a structural system as a whole (not only on a singular component) according to both objective and subjective parameters, that is, both load distribution parameters and spatial parameters, which are design related. The examined structural element was a canopy, situated in a dynamic environment that brought a change in the element’s load distribution. The learning ability was given by applying a supervised learning algorithm—Artificial Neural Network (ANN)—on a physical prototype. The ANN was trained by an optimized database of finite solutions, which was created by a Genetic Algorithm. Through this method, complex calculations are conducted “offline”, and the component operates in a “decision-making” mode in real time, adapting to a versatile environment while using minimal computational resources. Results show that the case study successfully exhibited self-learning and acquired the ability to adapt to unpredictable changing forces while keeping certain design requirements. This method can be applied over different structural elements (façade elements, canopies, structural components, etc.) to achieve adaptation to various parameters with an unpredictable pattern, such as human behavior or weather conditions.