Artificial Neural Network Learning Algorithms Research Articles

Explainable long-term building energy consumption prediction using QLattice

The global building sector is responsible for nearly 40% of total carbon emissions, offering great potential to move closer to set climate goals. Energy performance certificates designed to increase the energy efficiency of buildings require accurate predictions of building energy performance. With significant advances in information and communication technology, data-driven methods have been introduced into building energy performance research demonstrating high computational efficiency and prediction performance. However, most studies focus on prediction performance without considering the potential of explainable artificial intelligence. To bridge this gap, the novel QLattice algorithm, designed to satisfy both aspects, is applied to a dataset of over 25,000 German residential buildings for predicting annual building energy performance. The prediction performance, computation time, and explainability of the QLattice is compared to the established machine learning algorithms artificial neural network, support vector regression, extreme gradient boosting, and multiple-linear regression in a case study, variable importance analyzed, and appropriate applications proposed. The results show quite strongly that the QLattice should be further considered in the research of energy performance certificates and may be a potential alternative to established machine learning algorithms for other prediction tasks in energy research.

Comparative Analysis of Machine Learning Methods for Predicting Robotized Incremental Metal Sheet Forming Force.

This paper proposes a method for extracting information from the parameters of a single point incremental forming (SPIF) process. The measurement of the forming force using this technology helps to avoid failures, identify optimal processes, and to implement routine control. Since forming forces are also dependent on the friction between the tool and the sheet metal, an innovative solution has been proposed to actively control the friction forces by modulating the vibrations that replace the environmentally unfriendly lubrication of contact surfaces. This study focuses on the influence of mechanical properties, process parameters and sheet thickness on the maximum forming force. Artificial Neural Network (ANN) and different machine learning (ML) algorithms have been applied to develop an efficient force prediction model. The predicted forces agreed reasonably well with the experimental results. Assuming that the variability of each input function is characterized by a normal distribution, sampling data were generated. The applicability of the models in an industrial environment is due to their relatively high performance and the ability to balance model bias and variance. The results indicate that ANN and Gaussian process regression (GPR) have been identified as the most efficient methods for developing forming force prediction models.

Effective implementation of machine learning algorithms using 3D colour texture feature for traffic sign detection for smart cities

AbstractIn recent past, emerging technology, such as AI has revolutionized progress and advancement of traffic management solutions in context of smart cities. This paper describes a novel approach of using 3D colour‐texture based feature for image detection on public datasets of Chinese traffic sign research database (TSRD) and Mapillary image database. The implementation of 3D colour texture feature for traffic sign detection is evaluated using artificial neural network and multiple other machine learning algorithms as classifier for image detection purposes. Both datasets used in our experiments are considered among most diverse traffic signage datasets globally with annotations of almost all classes. Image datasets have been publicly made available to researchers for academic purposes. For classification of traffic sign images on Chinese TSRD and Mapillary datasets, the result outcome from back‐propagation neural network (NN) classifier outperforms all other and produced best results and under multiple ML algorithms, support vector machines (SVM) cubic has best results. For classification of traffic sign images on Mapillary dataset, the result outcome from rational quadratic Gaussian process regression has best results. Also, the comparisons of results with other similar novel works have also been discussed. The proposed approach outperformed the previous experiments as observed through comparative analysis performed on both datasets. The conclusion of the research work highlights that the image detection performance is noticeably improved by using the combined 3D colour‐texture feature based proposed approach.

A deep learning based decision support system for diagnosis of Temporomandibular joint disorder

Temporomandibular Joint sounds are a very common disorder in the general population. Temporomandibular Disorder (TMD) is any discomfort related to Temporomandibular Joint (TMJ). In this paper, a novel decision support system based on deep learning and neural network algorithms for diagnosis of Temporomandibular Joint disorder is introduced. A non-invasive device is designed for the recording of TMJ sounds. An interface is developed that will facilitate the dentist to operate on the recorded audio data. The collected data consist of the patient's left and right Temporomandibular Joint sound, ambient noise sound, the patient's clinical data, the dentist's notes about the patient, diagnosis, and treatment. Then signal processing, artificial neural network and deep learning algorithms are used to classify these measurements, and thus, the method that decides about the patient's condition is developed. Frequency, statistical and deep learning-based methods are compared in terms of classification success. The results show that the classification success of the classification method based on deep learning is consistently over 94.5% and it is more successful than the previous two methods. The proposed system can give the physician an idea about the effectiveness of the treatment methods applied to the patient in order to treat joint sounds which are among the important symptoms of TMD.

Discrimination of earthquakes and quarries in the Edirne district (Turkey) and its vicinity by using a linear discriminate function method and artificial neural networks

In this study, seismic events in the Edirne district (Turkey) and its vicinity have been investigated in order to discriminate earthquakes from quarry blasts. A total of 150 seismic events with Md ≤ 3.5 duration magnitude from a seismic activity catalog between 2009 and 2014 recorded by the Enez (ENEZ), Erikli (ERIK) and Gelibolu (GELI) broadband stations operated by Bogazici University, Kandilli Observatory and Earthquake Research Institute Regional Earthquake-Tsunami Monitoring Center were used in this study. The maximum S-wave and maximum P-wave amplitude ratio of vertical component velocity seismograms, power ratio (Complexity) and total signal duration of the waveform were calculated. Earthquakes and quarry blasts were discriminated using the linear discriminate function (LDF) and back propagation feed forward neural networks, an artificial neural network (ANN) learning algorithm, taking the determination coefficient and variance account values between these parameters into consideration. Eighty-one (54%) of the total 150 seismic events studied were determined to be earthquakes, and sixty-nine (46%) of them were determined to be quarry blasts. The LDF and ANNs methods were applied to the data in Edirne and its vicinity using a pair of parameters and were compared to each other for the first time. The accuracy of the methods are 95% and 99% for LDF and ANNs, respectively.

Box Office Success Prediction Through Artificial Neural Network and Machine Learning Algorithm

In this study, a comparison is made between Artificial Neural Network Algorithm and Machine Learning Algorithm of Support Vector Machine by considering pre-release and pre plus post release features of a movie for predicting a movie to a particular success class. These success classes includes Flop/Disaster, Average, Hit, Super-Hit and Block-Buster of a movie. Our best performing model from Artificial Neural Network and Support Vector Machine is able to give better performance in order to predict any particular movie to specific success class of a movie for both the cases of prerelease as well as pre plus post release features of a movie.

Performance of a language identification system using hybrid features and ANN learning algorithms

This paper discusses the use of hybrid features and artificial neural network (ANN) for spoken language identification (LID). At the feature extraction stage, we have used RASTA-PLP features and later hybrid features by combining the state-of-the-art MFCC features with RASTA-PLP features. The classifier used at the back-end of the LID system is Feed-forward Back Propagation Neural Network (FFBPNN). Performance comparison is done for all the thirteen learning algorithms available in FFBPNN. Results indicate better performance with MFCC + RASTA-PLP features as compared to RASTA-PLP features when used individually. The classification accuracy of 94.6 percent is obtained with ‘trainlm’ network training function and a test error rate of 0.10 with MFCC + RASTA-PLP hybrid features. On the other hand, RASTA-PLP features provide the best classification accuracy of 89.6%. Also, in this paper two new training functions are proposed. Experimental results show that for the MFCC + RASTA-PLP features, the proposed training function 1 offers 95.1 percent accuracy and 0.0993 mean square error. Similarly, the proposed training function 2 offers 95.3 per cent classification accuracy and 0.093 minimum mean square error for MFCC + RASTA-PLP features. The feasibility of the proposed solution is estimated by simulating multiple experiments on a language database of four languages i.e. Hindi, Tamil, Malayalam and English in the working Platform of MATLAB.

(Invited) Density Functional Theory Modeling - Machine Learning Approach for Redox Potential of Organic Electrode Materials

In this talk, we present the first-principles density functional theory modeling of redox properties of various carbon materials, including quinone derivatives, dopamine, and DNA. First, we discuss the contribution of electron affinity of molecules and solvation to redox potential from our density functional theory (DFT) calculations. Second, we discuss the relationship of redox potential with the design of molecular structure. It is found that aromaticity and electron-withdrawing functional groups can be used to tune the redox potential. Especially, the variation of redox potential as a function of Li association is discussed. The conclusion is that we can tune up the redox potential by designing the detailed molecular structure. Next, all these results are used to build up machine learning. We discuss our artificial neural network and other machine learning algorithms regarding its validation and predictability. Through this study, the redox characteristics of carbon materials are systematically investigated using the first-principles computational modeling method, and then utilized via machine learning methods for high-throughput screening process.

A Novel Assisted Artificial Neural Network Modeling Approach for Improved Accuracy Using Small Datasets: Application in Residual Strength Evaluation of Panels with Multiple Site Damage Cracks

An artificial neural network (ANN) extracts knowledge from a training dataset and uses this acquired knowledge to forecast outputs for any new set of inputs. When the input/output relations are complex and highly non-linear, the ANN needs a relatively large training dataset (hundreds of data points) to capture these relations adequately. This paper introduces a novel assisted-ANN modeling approach that enables the development of ANNs using small datasets, while maintaining high prediction accuracy. This approach uses parameters that are obtained using the known input/output relations (partial or full relations). These so called assistance parameters are included as ANN inputs in addition to the traditional direct independent inputs. The proposed assisted approach is applied for predicting the residual strength of panels with multiple site damage (MSD) cracks. Different assistance levels (four levels) and different training dataset sizes (from 75 down to 22 data points) are investigated, and the results are compared to the traditional approach. The results show that the assisted approach helps in achieving high predictions’ accuracy (<3% average error). The relative accuracy improvement is higher (up to 46%) for ANN learning algorithms that give lower prediction accuracy. Also, the relative accuracy improvement becomes more significant (up to 38%) for smaller dataset sizes.

Turkish handwriting recognition system using multi-layer perceptron

Recently, handwriting recognition has found many application areas along with technological advances. Handwriting recognition systems can greatly simplify human life by reading tax returns, forwarding mail, reading bank checks, and so on. On the other hand, these systems can reduce the need for human interaction. Therefore, academic and commercial studies of handwriting characters have recently become an important research topic in pattern recognition. In this study, Turkish handwritten letter recognition system from A to Z was developed in C++ environment by using Artificial Neural Networks (ANNs). After the feature data were extracted, handwriting images were presented to the network, the training process of ANN was completed, and different handwriting images were classified with trained ANN. In this study, MLP (Multi-Layered Perceptron: MLP) type ANN and back-propagation learning algorithm were used. The ANN used has 35 inputs and 23 outputs. In the hidden layer, ANNs with different numbers of artificial neural cells (neurons) were evaluated and the most appropriate neural number ANN was selected. As a result, ANN with 24 neurons was selected in the hidden layer and handwriting images was classified with an accuracy rate of 94.90 %.

A Bilevel Learning Model and Algorithm for Self-Organizing Feed-Forward Neural Networks for Pattern Classification.

Conventional artificial neural network (ANN) learning algorithms for classification tasks, either derivative-based optimization algorithms or derivative-free optimization algorithms work by training ANN first (or training and validating ANN) and then testing ANN, which are a two-stage and one-pass learning mechanism. Thus, this learning mechanism may not guarantee the generalization ability of a trained ANN. In this article, a novel bilevel learning model is constructed for self-organizing feed-forward neural network (FFNN), in which the training and testing processes are integrated into a unified framework. In this bilevel model, the upper level optimization problem is built for testing error on testing data set and network architecture based on network complexity, whereas the lower level optimization problem is constructed for network weights based on training error on training data set. For the bilevel framework, an interactive learning algorithm is proposed to optimize the architecture and weights of an FFNN with consideration of both training error and testing error. In this interactive learning algorithm, a hybrid binary particle swarm optimization (BPSO) taken as an upper level optimizer is used to self-organize network architecture, whereas the Levenberg-Marquardt (LM) algorithm as a lower level optimizer is utilized to optimize the connection weights of an FFNN. The bilevel learning model and algorithm have been tested on 20 benchmark classification problems. Experimental results demonstrate that the bilevel learning algorithm can significantly produce more compact FFNNs with more excellent generalization ability when compared with conventional learning algorithms.

Second Generation Current Conveyor Based Floating Fractional Order Memristance Simulator and a New Dynamical System

Abstract In recent years, due to its non-volatile memory, non-locality, and weak singularity features, fractional calculations have begun to take place frequently in artificial neural network implementations and learning algorithms. Therefore, there is a need for circuit element implementations providing fractional function behaviors for the physical realization of these neural networks. In this study, a previously defined integer order memristor element equation is changed and a fractional order memristor is given in a similar structure. By using the obtained mathematical equation, frequency-dependent pinched hysteresis loops are obtained. A memristance simulator circuit that provides the proposed mathematical relationship is proposed. Spice simulations of the circuit are run and it is seen that they are in good agreement with the theory. Also, the non-volatility feature has been demonstrated with Spice simulations. The proposed circuit can be realized by using the integrated circuit elements available on the market. With a small connection change, the proposed structure can be used to produce both positive and negative memristance values.

Multi-stage artificial neural network structure-based optimization of geothermal energy powered Kalina cycle

In this study, the geothermal energy powered Kalina cycle (GEP-KC) was optimized by using a multi-stage artificial neural network (ANN) analysis. The ANN model was basically composed of two stages. The first stage has one network, and the second stage consists of three networks in this ANN model. The 365 GEP-KCs were designed for four variable parameters. These designs were analytically analyzed by means of thermodynamic and economic analysis. The obtained data were used for modeling of multi-stage ANN structure. This multi-stage ANN model was designed with the aim of maximizing net present value (NPV). Turbine inlet pressure, geothermal water outlet temperature at evaporator, condenser pressure and ammonia mass fraction were input parameters of the multi-stage ANN model. Energy efficiency and exergy efficiency of the GEP-KC were outputs of the first stage, and the NPV of the GEP-KC was the output of the third network of the second stage. The most suitable network structure for the optimization of GEP-KC was performed by using the Levenberg–Marquardt variant of back-propagation learning algorithm for multi-stage ANN. The cov, MPE, RMSE and R2 values of multi-stage ANN were calculated as 2.558308, 1.077997189, 1.777658128 and 0.994693, respectively, for NPV. The calculated masses and biases of this structure were used to determine the optimum operating parameters of GEP-KC. The analytical findings of the NPV, energy and exergy efficiencies of the optimum GEP-KC model were, respectively, determined as 113.0732 M$, 6.7285% and 46.8701% in a high accuracy with the ANN results.

Developing Prediction System for Solar Power Plant Using Machine Learning Algorithms

The use of renewable energy sources in the production of electricity has become inevitable in order to reduce the greenhouse gases left in the atmosphere that cause the Earth to warm up. Although countries on a national basis have implemented a number of policies to support electricity generated from renewable energy sources, investments to produce electricity without a license on a local basis are not desirable. According to the climatic conditions of the power plant of 1 MW installed founded in Konya and power plant production data are monitored. Machine learning is a sub-branch of artificial intelligence that deals with the design and development of algorithms that allow computers to develop their behavior based on experimental data. In this study, Naive Bayes, Decision Tree, CN2 Rule Induction, Random Forest, Support Vector Machine, k-Nearest Neighbor, Artificial Neural Network, Logistic Regression and AdaBoost machine learning algorithms are used for prediction and classification. Generally, energy investors are curious about the return on their investment. It is very important for energy providers to predict how much electricity will be generated from existing solar power plants and accordingly determine the measures they will take to meet the electricity demand in the future. ROC analyzes were performed for machine learning models and performance evaluation was performed. In this study, the best performance estimation value obtained from the solar power plant depending on the weather conditions was obtained with 92.24% accuracy.

Research on Error Monitoring Model of CNC Machine Tool Based on Artificial Intelligence

In recent years, according to the rapid development and increasing application of artificial intelligence technology, the mechanical structure of modern CNC machine tools is becoming more and more complicated, and the automation level of CNC machine tools is constantly increasing. Research on a set of artificial intelligence-based CNC machine tool error monitoring model is of great significance to China’s industrial development. This article is a more in-depth and systematic study of the health monitoring system of CNC machine tools, especially high-end (such as five-axis linkage) CNC machine tools, multi-source fusion technology that supports high-reliability maintenance decisions, and key components of CNC machine tools based on artificial intelligence Fault monitoring and early warning. Based on the artificial intelligence SOM neural network learning algorithm, this paper designs a rolling bearing fault monitoring and early warning model architecture, studies the multi-source data fusion method based on DS evidence theory, and proposes and implements a self-organizing feature mapping network fault monitoring algorithm and adaptive ARMA The fault early warning algorithm improves the calculation efficiency of the rolling bearing monitoring diagnosis and early warning under the premise of ensuring accuracy. The experimental results show that when the sampling value is greater than 980 groups, the degradation value increases more until the bearing is in a failure state.

Deep Generative Models-Based Anomaly Detection for Spacecraft Control Systems.

A spacecraft attitude control system provides mechanical and electrical control to achieve the required functions under various mission scenarios. Although generally designed to be highly reliable, mission failure can occur if anomalies occur and the attitude control system fails to properly orient and stabilize the spacecraft. Because accessing spacecraft to directly repair such problems is usually infeasible, developing a continuous condition monitoring model is necessary to detect anomalies and respond accordingly. In this study, a method for detecting anomalies and characterizing failures for spacecraft attitude control systems is proposed. Herein, features are extracted from multidimensional time-series data of a simulation of the attitude control system. Then, the artificial neural network learning algorithms based on two types of generation models are applied. A Bayesian optimization algorithm with a Gaussian process is used to optimize the hyperparameters for the neural network to improve the performance. The performance is evaluated based on the reconstruction error through the algorithm using the newly generated data not used for learning as input data. Results show that the detection performance depends on the operating characteristics of each submode in the operation scenarios and type of generation model. The diagnostic results are monitored to detect anomalies in operation modes and scenarios.

Anomaly-Based Intrusion Detection System using Supervised Learning Algorithm Artificial Neural Network and Ant Colony Optimization with Feature Selection

In the advent of the cyber world, all know that cyber security is randomly used research area for researchers to secure host, network, and data because of increasingly complex attacks. In the advent of anomaly-based intrusion detection system, various techniques are applied to detect intrusion on system or network. This approach attains an extreme detection rate and accuracy but there may be overhead acquired to build and training them. The objective of this paper is to detect the intrusion of a system by proposing a Data mining technique which is based on supervised learning algorithm for training dataset. Artificial neural network (ANN) and Ant Colony Optimization (ACO) with feature selection are the basics of the proposed scheme. ACO work on a population-based algorithm and is motivated by the pheromone trail laying behavior of real ants, in which NSL-KDD Cup99 Dataset is used. Empirical Results clearly explain that the proposed system can attain an overall detection rate of 88% and time complexity of 0.343 sec, which is satisfactory when compared to other anomaly-based schemes.

Logic-Enhanced Adaptive Network-Based Fuzzy Classifier for Fall Recognition in Rehabilitation

Currently, the most widely adopted fall prediction methods use kinematic sensors for data collection and set thresholds to identify a fall based on artificial experience or neural network learning algorithms. These methods require classification and calibration in advance. Striking a balance between sensitivity and specificity is difficult when setting thresholds. Backpropagation (BP), support vector machines (SVMs) and other neural network algorithm models require professional expertise and are too complicated, difficult to understand, and computationally expensive, which renders embedded development with these models difficult. In response to these problems, this paper builds a fuzzy logic-enhanced adaptive neural network classifier (FLEANN); uses the prior knowledge of experts to construct fuzzy rules; employs neural networks to drive the underlying data to confirm and optimize logical rules; and utilizes an adaptive fuzzy neural network. The parameters of the algorithm are modified and optimized using a large amount of data training to improve the heuristics, transparency, and robustness of the neural network. The enhanced adaptive fuzzy neural network is applied in rehabilitation training for the classification of imprecise, uncertain, and nonlinear falls; abnormal gait; and normal gait. The experimental results confirm that the accuracy of this classifier is better than that of conventional neural network classifiers and that its logical structure is clear, which enables its application to embedded development. The reduction in computational demand enables the elimination of high-configuration PCs and reliance on a single chip microcomputer.

Implementation of ANN technique for performance prediction of solar thermal systems: A Comprehensive Review

Solar thermal systems (STS) are efficient and environmentally safe devices to meet the rapid increasing energy demand now a days. But it is very important to optimize their performance under required operating condition for efficient usage. Hence intelligent system-based techniques like artificial neural network (ANN) play an important role for system performance prediction in accurate and speedy way. In present paper, it is attempted to scrutinize the approach of ANN as an intelligent system-based method to accurately optimize the performance prediction of different solar thermal systems. Here, 25 research works related to various solar thermal systems have been reviewed and summarized to understand the impact of different ANN models and learning algorithms on performance prediction of STS. Using ANN, a brief stepwise summary of researchers’ work on various STS like solar air heaters, solar stills, solar cookers, solar dryers and solar hybrid systems, their predictions (results) and architectures (network and learning algorithms) in the literature till now, are also discussed here. This paper will genuinely help future researchers overview the work concisely related to solar thermal system performance prediction using various types of ANN models and learning algorithm and compare it with other global methods of machine learning. Citation: Ahmad, A., Ghritlahre, H. K., and Chandrakar, P. (2020). Implementation of ANN technique for performance prediction of solar thermal systems: A Comprehensive Review. Trends in Renewable Energy, 6, 12-36. DOI: 10.17737/tre.2020.6.1.00110

Assessment of Artificial Neural Networks Learning Algorithms and Training Datasets for Solar Photovoltaic Power Production Prediction

The capability of accurately predicting the Solar Photovoltaic (PV) power productions is crucial to effectively control and manage the electrical grid. In this regard, the objective of this work is to propose an efficient Artificial Neural Network (ANN) model in which 10 different learning algorithms (i.e., different in the way in which the adjustment on the ANN internal parameters is formulated to effectively map the inputs to the outputs) and 23 different training datasets (i.e., different combinations of the real-time weather variables and the PV power production data) are investigated for accurate one day-ahead power production predictions with short computational time. In particular, the correlations between different combinations of the historical wind speed, ambient temperature, global solar radiation, PV power productions, and the time stamp of the year are examined for developing an efficient solar PV power production prediction model. The investigation is carried out on a 231 kWac grid-connected solar PV system located in Jordan. An ANN that receives in input the whole historical weather variables and PV power productions, and the time stamp of the year accompanied with Levenberg-Marquardt (LM) learning algorithm is found to provide the most accurate predictions with less computational efforts. Specifically, an enhancement reaches up to 15%, 1%, and 5% for the Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Coefficient of Determination (R2) performance metrics, respectively, compared to the Persistence prediction model of literature.