Narrow Results

Volatility of gold price is of great significance for avoiding the risk of gold investment. It is necessary to understand the effect of external events and intrinsic regularities to make accurate price predictions. This paper first compared EMD with CEEMD algorithm, and the results find that CEEMD algorithm performance is better than that of EMD in analysis gold price volatility. Then this paper uses the complementary ensemble empirical mode decomposition (CEEMD) to decompose the historical price of international gold into price components at different frequencies, and extracts a short-term fluctuation, a shock from significant events and a long-term price. In addition, this paper combines the Iterative cumulative sum of squares (ICSS) with Chow test to test the three event prices for structural breaks, and analyzes the effect of external events on volatility of gold price by comparing the external events with the test results for structural breaks. Finally, this paper constructs support vector machine (SVM) models and artificial neural network (ANN) on three series for prediction, and finds that the SVM performed better in gold price prediction in one-step-ahead and five-step-ahead, and when we combine the SVMs and ANNs with price components to make predictions, the error of the combined prediction is smaller than SVMs and ANNs with separate terms of series extracted.

Equal channel angular pressing (ECAP) processed materials have higher grain refinement and strength, and they exhibit more surface roughness when they are machined. This enhancement in the properties highly influences the surface roughness and material removal rate of the materials. The commercial pure aluminum has a wide variety of applications when it is enhanced with high strength properties. In this paper, the machinability of commercially pure aluminum processed through ECAP is investigated in turning operations. Different ECAP processes are carried out to study the microstructural characterization and mechanical properties of the material. The material removal rate and surface roughness are tested by performing the turning operation in the CNC lathe with chemical vapor deposited carbide tool such that the feed rate, spindle speed and depth of cut are considered as the machining variables. To create a hypothesis for the experimentation, the empirical models are developed for the objective functions using the statistical technique response surface methodology (RSM) such that the response models are the objective functions and the model variables are the machining parameters. The response models are verified for the adequacy through ANOVA and $p$-test, and also verified for the closeness with the experimental results. Artificial neural network (ANN)-based empirical equations are also developed for the objective functions using the RSM design matrix and the results of both the RSM and ANN are compared for the suitability.

This paper aims to develop a predictive model and optimize the performance of the abrasive water jet machining (AWJM) during machining of carbon fiber-reinforced plastic (CFRP) epoxy laminates composite through a unique approach of artificial neural network (ANN) linked with the nondominated sorting genetic algorithm-II (NSGA-II). Initially, 80 AWJM experimental runs were carried out to generate the data set to train and test the ANN model. During the experimentation, the stand-off distance (SOD), water pressure, traverse speed and abrasive mass flow rate (AMFR) were selected as input AWJM variables and the average surface roughness and kerf width were considered as response variables. The established ANN model predicted the response variable with mean square error of 0.0027. Finally, the ANN coupled NSGA-II algorithm was applied to determine the optimum AWJM input parameters combinations based on multiple objectives.

In this study, a new prediction model is proposed to predict the 7-day compressive strength of ultrahigh-performance concrete (UHPC) with different mix proportions using artificial neural network (ANN) and support vector machine (SVM). The predicted results are compared with the experimental results to verify the proposed model. Then, the importance of each component and the sensitivity of parameters are investigated. The research proves that the proposed model can estimate the 7-day compressive strength of UHPC based on the mix proportions.

The implementation of the fused deposition modeling (FDM) technique in the production system is mainly due to its flexibility and ability to fabricate complex 3D prototypes and geometries. However, the mechanical strength of the printed parts needs to be investigated which was influenced by the process parameters such as layer thickness (LT), raster angle (RA), and Infill Density (ID). Therefore, these process parameters need to be optimized to attain better mechanical strength from the FDM printed parts. In this research, ePA-CF filament material was used to fabricate the specimens based on the selected process parameters such as LT (0.07, 0.14, and 0.20mm), RA (0^∘, 45^∘, and 90^∘) and ID (50%, 75%, and 100%). The artificial neural network (ANN) method was implemented to determine the influential printing process parameters. Tensile, flexural, and impact tests were considered as the response parameters based on the various combination of the input parameters. It was concluded that the printing of nylon carbon parts using $\mathrm{\text{LT}}=0.14$mm, $\mathrm{\text{ID}}=100\%$, $\mathrm{\text{RA}}=90^{\circ }$ retains improved tensile strength of 66 MPa, flexural strength of 87MPa and impact strength of 12.5KJ/m². Further, the propagation of cracks and the mode of failure were examined using SEM fractography. These observations substantiate that the selection of an optimal combination of FDM parameters assists in enhancing the mechanical strength of the printed nylon carbon parts.

High Carbon High Chromium (or AISI D2) Steels, owing to the fine surface finish they produce upon grinding, find lot of applications in die casting. Machining parameters affect the surface finish significantly during the grinding operation. In this context, this work puts an effort to arrive at the optimum machining parameters relating to fine surface finish with minimum cutting force. The material removal caused by the abrasive grinding wheel makes the process a very complex and nonlinear machining operation. In many situations, traditional optimization techniques fail to provide realistic optimum conditions because of the associated complexity. In order to overcome this issue, particle swarm optimization (PSO) coupled with artificial neural network (ANN) is applied in this research work for parameter optimization with the objective of achieving minimum surface roughness and cutting force. The machining parameters selected for the investigation were table speed, cross feed and depth of cut and the responses were surface roughness and cutting force. ANNs, inspired from biological neural networks, are well capable of providing patterns, which are too complex in behavior. The ANN model developed was used as the fitness function for PSO to complete the optimization. Optimization was also carried out using conventional response surface methodology-genetic algorithm (RSM-GA) approach in which regression equation developed with RSM was considered as the fitness function for GA. Confirmatory experiments were conducted and the comparison showed that PSO coupled with ANN is a reliable tool for complex optimization problems.

In this paper, a strategy has been set for minimizing the corner error in the modification of pulse and non-pulse parameters. Taguchi’s philosophy has been used to design the experiment by varying process parameters (i.e. Spark-on Time (STon), Wire Tension (WT), Servo-Voltage (Sv), Discharge Current (DC) and Wire-speed (Sw)), to explore the machining outcomes. The response characteristics have been measured in terms of cutting speed (Cs), Corner Error (CD) and surface Roughness (RA) using Topas plus X wire of $\varphi$ 0.25mm diameter. The machining performance characteristics were analyzed using main effect plots and analysis of variance (ANOVA). Furthermore, a soft computing-based hybrid optimization technique (Artifical Neural Network (ANN)-based Multi-Objective Grey Wolf Optimizer (MOGWO)) has been utilized to search the multi-optimum parameter setting for superior machining results. The most significant parameter observed is DC for Cs, which is determined to be 34.81%. Moreover, WT found 26.29% and 34.10% for CD and RA, respectively. The confirmation test shows that the maximum absolute percentage errors are observed as 3.89%, 6.3% and 9.7% for Cs, CD and RA, respectively. The proposed hybrid technique can generate superior solutions compared to the existing algorithms. Notably, the outcomes obtained on new instances exhibit potential, purposeful, and efficacy.

A Battery Management System (BMS) can prolong the life of the battery but it depends on the accuracy of the adopted scheme. Different techniques have been developed to enhance the BMS by monitoring the State of Health (SOH) of the battery. In this paper, the detection of battery voltage is analyzed by using the cycle counting method, which is a conventional technique and compared with Artificial Neural Network (ANN), a heuristic method. The advantage of the proposed ANN method is that SOH can be monitored without disconnecting the battery from the load. Also, the sampling data to the ANN are derived from various techniques including Open Circuit Voltage (OCV) method, Ambient temperature measurement, and valley point detection. A feed-forward backpropagation algorithm is used to achieve the purpose of real-time monitoring of the LAB. The results show that the precise estimation of SOH can be obtained by Feed-Forward Neural Network (FFNN) when trained with more sampling data.

The impact localization in composite panels is assessed using two machine learning techniques: least square support vector machines (LSSVM) and artificial neural networks (ANN) with local strain signals from piezoelectric sensors. Sensor signals from impact experiments on a composite plate as well as signals simulated by a finite element model are used to train and test models. A comparative study shows that LSSVM achieves better accuracy than ANN on identifying location of impacts for a combination of large mass impact and small mass impact, in particular when less data is available for training which is more appropriate for real aeronautical application. Additionally, LSSVM is more capable of identifying new impact events which have not been considered in the training process.

The BRICS (Brazil, Russia, India, China and South Africa) acronym was created by the International Monetary Foundation (IMF)–Group of Seven (G7) to represent the bloc of developing economies which crucially impact on the global economy by their potential economic growth. Most of the foreign direct investment are considering the stock markets of BRICS as the most attractive destination for foreign portfolio investment. This study aims to identify the relationship between macroeconomic variables and the stock market index values of BRICS and generate accurate predictions for index values by performing linear regression and artificial neural network hybrid models. Monthly data from January 2003 to December 2019 are used for the empirical study. The results indicate that a strong correlation exists between the stock market and macroeconomic variables in BRICS over time. The hybrid model is observed very accurate for index value prediction where the mean absolute percentage error (MAPE) value is 0.714% for the whole data set covering all BRICS countries data during the study period. Additionally, MAPE values for each of the BRICS countries are, respectively, obtained as 0.083%, 2.316%, 0.116%, 0.962% and 0.092%. Thus, the main findings of this study show that while neural network-integrated models have high performances for volatile stock market prediction, macroeconomic stabilization should be the priority of monetary policy to prevent the high volatility of stock markets.

The machining of metal matrix composites (MMCs) creates an extra challenge as compared to that of metals and alloys due to their hardness owing to the abrasive reinforcement particles. This paper presents the study on end milling of Al-4032/3% SiC composite considering the cutting speed (CS), feed rate (FR) and depth of cut (DOC) as the process parameters. Surface finish and material removal rate (MRR) have been taken as the response parameters. The Al-4032-based AMC has been prepared using stir casting process. Taguchi’s $L_{27}$ orthogonal array (OA) has been used for experimental trials. The optimum setting of the parameters has been obtained using TGRA. The resulting surface roughness (SR; $R_a$) occurs in the range of 1.18–3.97$\mu$m, with the minimum value corresponding to the CS of 110m/min, FR of 0.05mm/tooth and DOC of 1.2mm. Bayesian regularization (BR), scaled conjugate gradient (SCG) and Levenberg–Marquardt (LM) algorithms have been employed for training, validating and testing the 3–$n$–1 and 3–$n$–2 ANN architectures ($1\le n\le 20$). Minimum root-mean-square error (RMSE) has been taken as the standard for evaluating the model. Neural network toolbox of MATLAB “R2019A” has been used for prediction of the response.

Repetitive transcranial magnetic stimulation (rTMS) is defined as a noninvasive technique of brain stimulation conducted for both diagnostic and therapeutic purposes. rTMS can effectively excite the brain neurons and increase brain plasticity, which becomes particularly useful in psychiatric and neurological fields. Biomarkers that predict clinical outcomes in depression are essential for increasing the precision of treatments and clinical outcomes. The electroencephalogram (EEG) is a noninvasive neurophysiological test that is promising as a biomarker sensitive to treatment effects. The aim of our study was to investigate a novel nonlinear index of the resting state EEG activity as a predictor of clinical outcome and compare its predictive capacity to traditional frequency-based indices. EEG was recorded from 50 patients with treatment resistant depression (TRD) and 24 healthy comparison (HC) subjects. TRD patients were treated with excitatory rTMS to the dorsolateral prefrontal cortex (DLPFC) for 4–6 weeks. EEG signals were first decomposed using the ICA algorithm and the extracted components were then processed by time-frequency analysis. We then go on to compare the participants’ depression severity before, after, and 2 months after finishing the last treatment session using the proposed rTMS therapy. Absolute powers (APs), band powers (BPs), and theta and beta band entropies (BAs), which were extracted from the EEG, are used as features for the classification of changes in patients and normal cases after applying rTMS. Accordingly, we can go beyond the Beck score and clinically classify the EEG signal into two classes: depression and normal. The results demonstrated 78.37%, 74.32%, and 82.43% accuracy for artificial neural network (ANN), $k$-nearest neighbor (KNN), and support vector machine (SVM) classifiers, respectively, indicating the superiority of the proposed method to those mentioned in similar studies. Also, the electrophysiological changes are shown to be evident in patients with major depression. Our data show that the time-frequency index yields superior outcome prediction performance compared to the traditional frequency band indices. Our findings warrant further investigation of EEG-based biomarkers in depression.

Agriculture not only plays a vital role in human survival but also contributes to the nation’s greater economic development. With the use of technologies like IoT, WSNs, remote sensing, camera surveillance, and many more, precision agriculture is the newest buzzword in the field of technology. Its primary goal is to lessen the labour of farmers while increasing the output of farms. Many machine learning techniques are designed to improve the productivity and quality of the crops, but the improper irrigation and disease prediction of the existing techniques leads to loss of productivity and quality. Hence, the IoT based wireless communication system is designed for smart irrigation and rice leaf prediction using ANN and ResNeXt-50 model. In this designed model, smart irrigation is controlled by collecting the temperature and moisture of the soil in the agricultural field by using the WSN. Likewise, a surveillance camera is placed in the agricultural field to capture the rice leaf to find the disease such as rice blast, leaf smut, brown spot and bacterial blight. These collected data are processed and classified for smart irrigation and rice leaf disease prediction. For evaluating the performance of both the ANN and ResNeXt-50 trained model, the performance metrics such as accuracy, sensitivity, specificity, precision, error etc. The performance metrics for the ANN and ResNeXt-50 model are 0.9427, 0.925, 0.903, 0.86, 0.0573 and 0.967, 0.935, 0.943, 0.939 and 0.033. Thus, the evaluation of the designed model results that the proposed approach is performing better compared to the current techniques.

New advancements in deep learning issues, motivated by real-world use cases, frequently contribute to this growth. Still, it’s not easy to recognize the speaker’s emotions from what they want to say. The proposed technique combines a deep learning-based brain-inspired prediction-making artificial neural network (ANN) through social ski-driver (SSD) optimization techniques. When assessing speaker emotion recognition (SER), the recognition results are compared with the existing convolutional neural network (CNN) and long short-term memory (LSTM)-based emotion recognition methods. The proposed method for classification based on ANN decreases the computational costs. The SER algorithm allows for a more in-depth classification of different emotions because of its relationship to ANN and LSTM. The SER model is based on ANN and the recognition impact of the feature reduction. The SER in this proposed research work is based on the ANN emotion classification system. Speaker recognition accuracy values of 96.46%, recall values of 95.39%, precision values of 95.21%, and F-Score values of 96.10% are obtained in this proposed result, which is higher than the existing result. The average accuracy results by using the proposed ANN classification technique are 4.38% and 2.89%, better than the existing CNN and LSTM techniques, respectively. The average precision results by using the proposed ANN classification technique are 4.67% and 2.49%, better than the existing CNN and LSTM techniques, respectively. The average recall results by using the proposed ANN classification technique are 2.90% and 1.42%, better than the existing CNN and LSTM techniques, respectively. The average precision results using the proposed ANN classification technique are 3.80% and 3.10%, better than the existing CNN and LSTM techniques, respectively.

The main goal of this study is to investigate whether social media, as a recent communication channel, has an impact on customer lifetime value (CLV). No studies have been done in Turkey with similar purposes in the telecommunication sector. To reach this goal, there has been an attempt to develop both artificial neural network models and sector-specific applicable models. Four years of data between 2011 and 2014 belonging to customers in the telecommunication sector who have a Twitter account are used in this study. The CLV is modeled through radial basis function (RBF), multilayer perceptron (MLP), and Elman neural network approaches, and the performance of such models is compared. According to the findings, calculated CLV error values are at an acceptable range in all formed models. Additionally, it is determined that the CLV was calculated with a lower error value in models where social media variables were used. The Elman neural network is determined to perform better compared to RBF and MLP.

For machining a component, it is important to understand the characteristics of work material in order to choose the appropriate cutting tool and to fix a set of machining parameters to achieve optimum output. Analytical models of machining processes require complete understanding of process mechanism and hence are difficult to be developed. Once developed, these models are useful in parametric optimization, process simulation, operation and process planning, process parameter selection, parametric analysis, process performance prediction, verification of the experimental results, and improving the process performance by implementing/incorporating the theoretical findings.

Neural network models associated with artificial intelligence are known as artificial neural networks (ANNs) which are simple mathematical models in the form of defining a function. This work presents the details of the experiments carried out for data acquisition, method of building the ANN models and their validation. These models can be used for predicting the output for a chosen set of input variables or for a specific desired output, finding the set of input variables to be chosen.

This work resulted in developing models for the turning process for Inconel 718 alloy in a scientific manner. It also enables further scope of identifying the optimized set of turning parameters for Inconel 718 material using the newly developed coated carbide tools, achieving quality surface and productivity.

Electrical Discharge Machining (EDM) is a thermal energy based non-traditional shaping process for shaping of hard and brittle electrically conductive materials, but it suffers with low machinability and recast layer formations. The combination of grinding with EDM means enhancement in machining capability, but the process becomes highly complex. Therefore, the assortment of control factors for optimum results is greatly challenging for the industries. The objective of present study is to optimize the control factors such as current, pulse on-time, pulse off-time, wheel RPM and abrasive grit number (GN) to optimize the material removal rate (MRR) and average surface roughness (Ra) for Grinding Aided-EDM process. For this purpose, the simultaneous application of soft computing methods such as Artificial Neural Network (ANN) and Genetic Algorithm (GA) has been employed. The results demonstrate that combination of ANN with GA effectively predicts the data and provides optimal results with adequate percentage errors in MRR and Ra positively.

In computational fluid dynamics (CFD), there is a transformation of methods over the years for building commercially coded software. Each method has predicted its own set of importance, but the exportation and prediction of data are some of the crucial elements for post-processing and validating results. In the present investigation, a detailed comparative analysis is performed over finite difference method (FDM) and finite volume method (FVM) method for the 1D steady-state heat conduction problem over a 1-m-long plate. The comparison was made between solution creation and validation between FDM and FVM for the analytical and computational scheme. The convergence-dependent study is performed as multi-objective optimization to predict how artificial neural network (ANN) can be used to verify and validate the solution of CFD.

The heterogeneous fluid model is expressed for the nanofluid flow to study the significance of the Fourier’s and Fick’s laws. The magnetohydrodynamics couple-stress bioconvective nanofluid flow is considered across an extending surface with the effect of exponential heat source/sink and stratified boundary conditions. The solid nanoparticulates and concentrations of motile microorganisms are added to the nonlinear system of differential equations conveying the non-Newtonian nanoliquid flow model. Moreover, the combined effect of the heat flux and thermal radiation is also evaluated. The similarity transformations are employed to transfigure the system of partial differential equations into the lowest order of ordinary differential equations. The Artificial Neural Network Levenberg Marquardt Back-propagation optimization algorithm is employed to solve these equations. To authenticate the outcomes, the dataset is formed using the MATLAB package “bvp4c”. The dataset is created for diverse circumstances of flow factors, as well as validation and testing of the Artificial Neural Network. The accuracy of the model is estimated through numerous statistical tools (histogram, curve fitting, regression measures, and performance plots). The outcomes are presented through the table and figures. It has been noticed that the couple-stress nanofluid flow declines with the influence of magnetic field and mixed convection. The couple-stress nanofluid temperature augments with the enhancement of the thermophoresis effect, buoyancy ratio factor, Rayleigh number, and thermal radiation. Moreover, the concentration curve lessens under the impact of the Lewis number while enriched with the outcome of the concentration stratification parameter. The absolute error of reference and targeted date is attained within 10${}^{-3}$–10${}^{-6}$ that proves the exceptional precision of the results.

In this paper, an effort is made to determine the optimized parameters in laser welding of Hastelloy C-276 using Artificial Neural Network (ANN) and Genetic Algorithm (GA). CO₂ Laser welding was performed on a sheet of thickness 1.6mm based on Taguchi L27 orthogonal array. Laser power, welding speed and shielding gas flow rate were chosen as input parameters and Bead width, depth of Penetration and Microhardness were measured for assessing the weld quality. ANN was applied for modeling the welding process parameters i.e. heat input, welding speed and gas flow rate. Various learning algorithms such as Batch Back Propagation (BBP), Incremental Back Propagation (IBP), Quick Propagation (QP) and Levenberg–Marquardt (LM) were comprehensively tested for estimating the output parameters and a comparison was also made among them, with respect to prediction accuracy. BBP method was found to be the best learning algorithm. Experimental validation test was performed based on the ANN and GA predicted optimized responses and this welding input parameters provided satisfactory weld metal characteristics in terms of penetration depth, bead width and microhardness.