Are Artificial Neural Networks Research Articles

State-of-the-art review on the use of AI-enhanced computational mechanics in geotechnical engineering

Significant uncertainties can be found in the modelling of geotechnical materials. This can be attributed to the complex behaviour of soils and rocks amidst construction processes. Over the past decades, the field has increasingly embraced the application of artificial intelligence methodologies, thus recognising their suitability in forecasting non-linear relationships intrinsic to materials. This review offers a critical evaluation AI methodologies incorporated in computational mechanics for geotechnical engineering. The analysis categorises four pivotal areas: physical properties, mechanical properties, constitutive models, and other characteristics relevant to geotechnical materials. Among the various methodologies analysed, ANNs stand out as the most commonly used strategy, while other methods such as SVMs, LSTMs, and CNNs also see a significant level of application. The most widely used AI algorithms are Artificial Neural Networks (ANN), Random Forest (RF), and Support Vector Machines (SVM), representing 35%, 19%, and 17% respectively. The most extensive AI application is in the domain of mechanical properties, accounting for 59%, followed by other applications at 16%. The efficacy of AI applications is intrinsically linked to the type of datasets employed, the selected model input. This study also outlines future research directions emphasising the need to integrate physically guided and adaptive learning mechanisms to enhance the reliability and adaptability in addressing multi-scale and multi-physics coupled mechanics problems in geotechnics.

Artificial intelligence and machine learning in environmental impact prediction for soil pollution management – case for EIA process

Scientific predictions are a key component of Environmental Impact Assessments (EIA), which can indicate the level of change within an environmental sphere (e.g., soil). As part of the EIA process, decision-making in mitigating complex environmental problems such as maintaining soil quality can be challenging, especially in data-sparse locations. Artificial Intelligence (AI) can ameliorate but the literature suggests that the deployment of Machine Learning (ML) techniques in soil research is concentrated mostly in developed countries. The potential of ML in managing soil pollution from complex mixture of heavy metals, petroleum hydrocarbons, and physicochemical factors is rarely explored. To address this research gap, we built robust models that increase the accuracy of impact prediction based on new experimental soil data from a data-sparse region of Africa (i.e., Nigeria). The algorithms applied are artificial neural networks (ANN), support vector regression (SVR), regression tree (RT), and random forest (RF). The study also implemented a multivariate linear regression (MLR) model as a baseline. Key findings include (a) the MLR model performed less than the machine learning models largely due to the nonlinearity of data; (b) Log-normalization helped to improve the predictive capability of all models as the effects of statistical variability were removed; (c) the RF model had the best performance in terms of correlation coefficient, mean absolute error, and root mean square error, and (d) the machine learning models showed improved performance with increased correlation and lower error between the actual and predicted soil electrical conductivity values. Our results imply that data sparsity may no longer be an excuse for the non-use of quantitative impact prediction in Environmental Impact Assessment (EIA) processes. This could change how EIAs are conducted and enhance sustainability in natural resource exploitation, globally. Future work will apply algorithms for automated feature selection to obtain optimal subset of soil quality measurements that will further improve the accuracy of the models.

Comparison of Machine Learning Algorithms for Mass Appraisal of Real Estate Data

Abstract In recent years, machine learning algorithms have been used in the mass appraisal of real estate. In this study, 5 machine learning algorithms are used for residential type real estate. Machine learning algorithms used for mass appraisal in this study are Artificial Neural Networks (ANN), Random Forest (RO), Multiple Regression Analysis (MRA), K-Nearest Neighborhood (k-nn), Support Vector Regression (SVR). To test the study, real estate data collected from the central districts of Ankara, were used. The main purpose of this study is to find out which machine learning algorithm gives the best results for the mass appraisal of real estates and to reveal the most important variables that affect the prices of real estate. According to the results obtained for the city of Ankara, it was observed that the best algorithm for mass appraisal is RF in residential-type real estates, followed by the ANN, k-nn, and linear regression algorithms, respectively. According to the results obtained from the residential real estate, it was concluded that heating and distances to places of importance had the greatest effect on the value.

A Review on Prediction of Heart Disease based on Machine Learning and Datamining Techniques

Heart is the important organ in human body which supplies blood to all organs of the body. The abnormal situation of heart is considered as heart disease. According to WHO data cardiovascular, respiratory and neonatal conditions are the top three causes of Deaths in the World. In the year 2019 Heart diseases occupies 16%(9 million) of overall deaths happened in World. From two decades there is 4 times increase in the deaths with heart diseases this is because of change in life style, lack of physical activity, food habits, obesity ,stress, cholesterol ,high blood pressure and diabetes.so there is a need to work on prediction of heart diseases to save many lives because prediction is the only way to prevent the disease. In this paper we will discuss about existing algorithms and existing work done in different machine learning and datamining techniques, which are concentrated more on the classification and prediction. main objective is to evaluate the performance of these algorithms and identify the most accurate and efficient approach for diagnosing heart diseases.
 Some of the machine learning and data mining techniques are Artificial Neural Network(ANN),Decision Tree, Naive Bayes, SVM(Support Vector Machine),k-Nearest Neighbours (KNN),J48,SMO,Random forest and classification Tree.

Performance Analysis of Different Machine Learning Algorithms on Credit Card Fraud Detection

Machine learning (ML) is a logical investigation of various algorithms and factual models that PCs utilize to carry out particular operations that are not clearly programmed. This paper aims to statistically analyze different machine learning algorithms, and compare and contrast their performance for credit card fraud detection. Algorithms used are Artificial Neural Networks(ANN), Sample Vector Machine (SVM), and Kth Nearest Neighbour (KNN), Decision Tree, Logistic Regression and Random Forest. All these above mentioned algorithms are compared on basis of performance measures. It is deduced that the random forest algorithm is the best algorithm.

A neural network approach for classification of fault-slip data in geoscience

In geoscience, paleostress studies are a vital tool for understanding the tectonic evolution of the region. The collected hundreds or even thousands of heterogeneous fault-slip data need to be divided into homogeneous (i.e., belonging to similar tectonic environments) subgroups by geologists. Computer-based paleostress inversion programs are able to run homogenous data sets. Here, we are aiming for the classification of heterogeneous fault-slip data into homogeneous sub-data which is performed by using several techniques in Machine Learning (ML) algorithms, which are Artificial Neural Network (ANN), Naïve Bayes (NB) and Logistic Regression (LR) models. When these models are executed on the Anaconda Navigator interface with python language, the accuracies are obtained as 87.17 % for ANN, 79.71 % LR, and 62.1 % NB.

Quality and Defect Prediction in Plastic Injection Molding using Machine Learning Algorithms based Gating Systems and Its Mathematical Models

To achieve high quality products from Plastic Injection Molding (PIM) process it is very essential to identify the defective operations in automatic manner which is most challenging task. This paper proposes a Machine Learning (ML) approach to detect the complex faults occurrence during the PIM process. During initial sampling process of molding to achieve high quality and low time consumption it is essential to concentrate on the suitable determination of parameter values by considering the properties of injection molding process. For that purpose, a novel machine learning algorithms based gating system is introduced in PIM (MLGS-PIM). Technical evaluation can be done using simulation which combines the CATIA and MATLAB. Therefore in MLGS-PIM, a holistic approach is introduced to improve and predict the process quality of the parameters which is based on machine learning approaches. The considered machine learning approaches for this process are Artificial Neural Network (ANN) and Support Vector Machine (SVM). This two learning models are combined to achieve high quality under various conditions. Such novel ML based technique helps to increase the quality characteristics of the injection molding process and it is predicted with various parameter values where the simulation data and measurements are handled in an intelligent manner. The materials which are considered in the PIM process are thermoplastic polystyrene, thermoplastic acrylonitrile butadiene styrene and thermoplastic polyvinyl chloride where three types are gating systems are applied with it and consists of 3, 4 and 5 gates and as well the parameters which are measured for the output analysis are sum rate, bit error rate and convergence plot. The results show that the performance of the proposed MLGS-PIM approach significantly increases the performance when compared with the earlier approaches such as AntLion Optimization and PSO-MSQPA.

Artificial-Intelligence-Based Models Coupled with Correspondence Analysis Visualization on ART-Cases from Gombe State, Nigeria: A Comparative Study.

Antiretroviral therapy (ART) is the common hope for HIV/AIDS-treated patients. Total commitments from individuals and the entire community are the major challenges faced during treatment. This study investigated the progress of ART in the Federal Teaching Hospital in Gombe state, Nigeria by using various records of patients receiving treatment in the ART hospital unit. We combined artificial intelligence (AI)-based models and correspondence analysis (CA) techniques to predict and visualize the progress of ART from the beginning to the end. The AI models employed are artificial neural networks (ANNs), adaptive neuro-fuzzy inference systems (ANFISs) and support-vector machines (SVMs) and a classical linear regression model of multiple linear regression (MLR). According to the outcome of this study, ANFIS in both training and testing outperformed the remaining models given the R2 (0.903 and 0.904) and MSE (7.961 and 3.751) values, revealing that any increase in the number of years of taking ART medication will provide HIV/AIDS-treated patients with safer and elongated lives. The contingency results for the CA and the chi-square test did an excellent job of capturing and visualizing the patients on medication, which gave similar results in return, revealing there is a significant association between ART drugs and the age group, while the association between ART drugs and marital status (93.7%) explained a higher percentage of variation compared with the remaining variables.

Prediction of Laminar Burning Speed of Propane/Hydrogen/Air Mixtures Using Power-Law Correlation and Two Machine Learning Models

Abstract Propane (C3H8) and hydrogen (H2) are regarded as alternative fuels that are favorable to the environment. Hydrogen gas's low energy density, storage, and transportation are the main issues with using it as an alternative fuel. Addition of hydrogen gas in the combustion of propane will also improve flame stability, broaden lean flammability limits, and reduces pollutant emissions. Thus, utilizing propane and hydrogen mixtures as fuel is a good choice. Laminar burning speed is a fundamental property of a combustible mixture and can be used to provide information regarding the mixture’s reactivity, exothermicity, and diffusivity. In this study, power-law correlation and machine learning methods were used to create models that predict the laminar burning speed of propane/hydrogen/air mixtures at various states. Two machine learning models are artificial neural network (ANN) and support vector machine (SVM). The data were generated by using CANTRA code and a chemical kinetic mechanism. For a wide variety of input values, the models were able to determine the laminar burning speed with great accuracy. The ANN model yields the best performance. The main advantage of these models is the noticeably faster computing time when compared to chemical reaction mechanisms.

New Trends in the Modeling of Diseases Through Computational Techniques

The computational techniques are a set of novel problem-solving methodologies that have attracted wider attention for their excellent performance. The handling strategies of real-world problems are artificial neural networks (ANN), evolutionary computing (EC), and many more. An estimated fifty thousand to ninety thousand new leishmaniasis cases occur annually, with only 25% to 45% reported to the World Health Organization (WHO). It remains one of the top parasitic diseases with outbreak and mortality potential. In 2020, more than ninety percent of new cases reported to World Health Organization (WHO) occurred in ten countries: Brazil, China, Ethiopia, Eritrea, India, Kenya, Somalia, South Sudan, Sudan, and Yemen. The transmission of visceral leishmaniasis is studied dynamically and numerically. The study included positivity, boundedness, equilibria, reproduction number, and local stability of the model in the dynamical analysis. Some detailed methods like Runge Kutta and Euler depend on time steps and violate the physical relevance of the disease. They produce negative and unbounded results, so in disease dynamics, such developments have no biological significance; in other words, these results are meaningless. But the implicit nonstandard finite difference method does not depend on time step, positive, bounded, dynamic and consistent. All the computational techniques and their results were compared using computer simulations.

Estimation of density log and sonic log using artificial intelligence: an example from the Perth Basin, Australia

It is well understood that with a large number of data, an excellent interpretation of the subsurface condition can be produced, and also our understandings of the subsurface conditions can be improved significantly. However, having abundant subsurface geological and petrophysical data sometimes may not be possible, mainly due to budget issues. This situation can generate issues during hydrocarbon exploration and/or development activities.
 In this paper, the authors tried to apply artificial intelligence (AI) techniques to estimate outcomes values of particular wireline log data, using available petrophysic data. Two types of AI were selected and these are artificial neural network (ANN), and multiple linear regression (MLR). This research aims to advance our understanding of AI and its application in geology. There are three objectives of this study: (1) to estimate sonic log (DT) and density log (RhoB) using different types of AI (ANN and MLR); (2) to assess the best AI technique that can be used to estimate certain wireline log data; and (3) to compare the estimated wireline log values with the real, recorded values from the subsurface.
 Findings from this study show that ANN consistently provided a better accuracy percentage compared to MLR when estimating density log (RhoB). While using different set of data and technique, estimation of sonic log (DT) produced different accuracy level. Moreover, crossplot validation of the results show that the results from ANN analysis produced higher trendline reliability (R2) and correlation coefficient (R) than the results from MLR analysis. Comparison of the estimated RhoB and DT log data with the original recorded data shows minor mismatch. This is evident that AI technique can be a reliable solution to estimate particular outcomes of wireline log data, due to limited availability of the original recorded subsurface petrophysic data. It is expected that these findings would provide new insights into the application of AI in geology, and encourage the readers to explore and expand the many possibilities of the application of AI in geology.

On the interpretability of machine learning regression for path-loss prediction of millimeter-wave links

Millimeter-wave communication systems design requires accurate path-loss (PL) prediction, which is critical to determining coverage area and system capacity. In this work, four machine learning (ML) algorithms are proposed for PL prediction in an indoor environment for 5G millimeter-wave frequencies, from 26.5 to 40 GHz. They are artificial neural network (ANN), support vector regression (SVR), random forest (RF), and gradient tree boosting (GTB). We compare their performances, including also the empirical PL models ABG and CIF incorporating the number of crossed walls, as we propose. For the ML models, we present a methodology to select the predictor coalition by examining the marginal performance and the interpretability gains. The methodology allows choosing the most appropriate predictor coalition for building machines while not overlooking the non-linear connections between predictors and the path-loss. The results show that the ML techniques improve the prediction accuracy of empirical models. Among the ML algorithms employed, considering the optimal subset of predictors, the GTB model presents the best prediction and generalization capacities.

Sağlıklı ve Hasta Etlik Piliçlerin Kuş Gribi Açısından Yapay Sinir Ağları Kullanarak Otomatik Sınıflandırılması

Poultry meat containing low fat and high protein is an important and economical protein source in providing the animal protein requirement for human nutrition. The frequent emergence of poultry diseases such as avian influenza is the feature of fast-spread in farms seriously threatens both the economy and human health. In this study, neural network (NNs) models are proposed for the classification of broiler chickens as healthy and sick for earlier detection of poultry diseases. The NNs used in the classification are artificial neural network (ANN), adaptive neuro-fuzzy inference system (ANFIS), and support vector machine (SVM). In the literature, the data set which includes seven visual features were acquired through the IPTs and was used for training, testing, and validating the process of NN models. The results point out that, the computer vision-based application using NNs is successfully classified the broilers in terms of their health conditions.

An Integrated Classification and Association Rule Technique for Early-Stage Diabetes Risk Prediction.

The number of diabetic patients is increasing yearly worldwide, requiring the need for a quick intervention to help these people. Mortality rates are higher for diabetic patients with other serious health complications. Thus, early prediction for such diseases positively impacts healthcare quality and can prevent serious health complications later. This paper constructs an efficient prediction system for predicting diabetes in its early stage. The proposed system starts with a Local Outlier Factor (LOF)-based outlier detection technique to detect outlier data. A Balanced Bagging Classifier (BBC) technique is used to balance data distribution. Finally, integration between association rules and classification algorithms is used to develop a prediction model based on real data. Four classification algorithms were utilized in addition to an a priori algorithm that discovered relationships between various factors. The named algorithms are Artificial Neural Network (ANN), Decision Trees (DT), Support Vector Machines (SVM), and K Nearest Neighbor (KNN) for data classification. Results revealed that KNN provided the highest accuracy of 97.36% compared to the other applied algorithms. An a priori algorithm extracted association rules based on the Lift matrix. Four association rules from 12 attributes with the highest correlation and information gain scores relative to the class attribute were produced.

PERBANDINGAN ALGORITMA SUPERVISED MACHINE LEARNING UNTUK SISTEM PENGHINDARAN HALANGAN PADA ROBOT ASSISTANT UDAYANA 02 (RATNA02)

Supervised Machine Learning can make robots smarter by making decisions automatically. This study compares various Supervised Machine Learning algorithms to determine the best algorithm that can be used on Robot Assistant Udayana 02 (RATNA02). The algorithms to be compared are Artificial Neural Network (ANN), Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), Naïve Bayesian (NB), and K-Nearest Neighbor (KNN). Models were created using the TensorFlow and SKLearn libraries. The model is trained using 100.000 data of left sensor, right sensor, front sensor, robot offset, and data label. Data preprocessing is done using MinMaxScalar and LabelEncoder. The comparisons that will be measured are accuracy, training duration, and file size of the model. DT and RF algorithms get 100% accuracy followed by ANN, KNN and SVM with 99.87%, 97.42% and 98.52% respectively, and NB with 87.34%. Fastest training duration was achieved by NB for 0.03 seconds, followed by DT for 0.08 seconds, while other algorithms took more than one second. The smallest file size is owned by NB with a size of 2kb and DT ranks second with 4kb, other algorithms have a file size of more than 25kb. Decision Tree Algorithm is the best because the duration of the model training is relatively fast, the file size is small, and the accuracy is high.

Implementation of Soft Computing Techniques in Predicting and Optimizing the Operating Parameters of Compression Ignition Diesel Engines: State-of-the-Art Review, Challenges, and Future Outlook

Abstract Fossil fuels being the primary source of energy to global industrialization and rapid development are being consumed at an alarming rate, thus creating a dire need to search for alternative fuels and optimize the internal combustion (IC) engine performance parameters. Traditional methods of testing and optimizing the performances of IC engines are complex, time consuming, and expensive. This has led the researchers to shift their focus to faster and inexpensive techniques like soft computing (SC), which predict the optimum performance with a substantial accuracy. The SC techniques commonly used are artificial neural network (ANN), fuzzy logic, adaptive neuro-fuzzy inference system (ANFIS), genetic algorithm (GA), particle swarm optimization (PSO), and hybrid techniques like ANN-GA, ANN-PSO, and others. The data of engine parameters predicted with these models have been found to be in very close indices with the experimented values making them a reliable predicting tool. The ANN, fuzzy logic, and ANFIS models have been found to have a correlation coefficient (R) above 0.9 suggesting a good level of agreement between experimented and predicted values of several engine-out parameters. In the present review article, the application of various SC techniques in the prediction and the optimization of output parameters of compression ignition (CI) diesel engines are thoroughly reviewed along with their future prospects and challenges. This review work highlights the implication of these SC techniques in CI diesel engines run on both conventional fuel as well as biodiesels.

Local Texture Representation for Timber Defect Recognition based on Variation of LBP

This paper evaluates timber defect classification performance across four various Local Binary Patterns (LBP). The light and heavy timber used in the study are Rubberwood, KSK, Merbau, and Meranti, and eight natural timber defects involved; bark pocket, blue stain, borer holes, brown stain, knot, rot, split, and wane. A series of LBP feature sets were created by employing the Basic LBP, Rotation Invariant LBP, Uniform LBP, and Rotation Invariant Uniform LBP in a phase of feature extraction procedures. Several common classifiers were used to further separate the timber defect classes, which are Artificial Neural Network (ANN), J48 Decision Tree (J48), and K-Nearest Neighbor (KNN). Uniform LBP with ANN classifier provides the best performance at 63.4%, superior to all other LBP types. Features from Merbau provide the greatest F-measure when comparing the performance of the ANN classifier with Uniform LBP across timber fault classes and clean wood, surpassing other feature sets.

A System to automate the development of anomaly-based network intrusion detection model

Cyber security is the major concern in today’s world. Over the past couple of decades, the internet has grown to such an extent that almost every individual living on this planet has the access to the internet today. This can be viewed as one of the major achievements in the human race, but on the flip side of the coin, this gave rise to a lot of security issues for every individual or the company that is accessing the web through the internet. Hackers have become active and are always monitoring the networks to grab every possible opportunity to attack a system and make the best fortune out of its vulnerabilities. To safeguard people’s and organization’s privacy in this cyberspace, different network intrusion detection systems have been developed to detect the hacker’s presence in the networks. These systems fall under signature based and anomaly based intrusion detection systems. This paper deals with using anomaly based intrusion detection technique to develop an automation system to both train and test supervised machine learning models, which is developed to classify real time network traffic as to whether it is malicious or not. Currently the best models by considering both detection success rate and the false positives rate are Artificial Neural Networks(ANN) followed by Support Vector Machines(SVM). In this paper, it is verified that Artificial Neural Network (ANN) based machine learning with wrapper feature selection outperforms support vector machine (SVM) technique while classifying network traffic as harmful or harmless. Initially to evaluate the performance of the system, NSL-KDD dataset is used to train and test the SVM and ANN models and finally classify real time network traffic using these models. This system can be used to carry out model building automatically on the new datasets and also for classifying the behaviour of the provided dataset without having to code.

Revealing Unknown Controlled Substances and New Psychoactive Substances Using High-Resolution LC-MS-MS Machine Learning Models and the Hybrid Similarity Search Algorithm.

High-resolution liquid chromatography-tandem mass spectrometry (LC--MS-MS)-based machine learning models are constructed to address the analytical challenge of identifying unknown controlled substances and new psychoactive substances (NPSs). Using a training set composed of 770 LC-MS-MS barcode spectra (with binary entries 0 or 1) obtained generally by high-resolution mass spectrometers, three classification machine learning models were generated and evaluated. The three models are artificial neural network (ANN), support vector machine (SVM) and k-nearest neighbor (k-NN) models. In these models, controlled substances and NPSs were classified into 13 subgroups (benzylpiperazine, opiate, benzodiazepine, amphetamine, cocaine, methcathinone, classical cannabinoid, fentanyl, 2C series, indazole carbonyl compound, indole carbonyl compound, phencyclidine and others). Using 193 LC-MS-MS barcode spectra as an external test set, accuracy of the ANN, SVM and k-NN models were evaluated as 72.5%, 90.0% and 94.3%, respectively. Also, the hybrid similarity search (HSS) algorithm was evaluated to examine whether this algorithm can successfully identify unknown controlled substances and NPSs whose data are unavailable in the database. When only 24 representative LC-MS-MS spectra of controlled substances and NPSs were selectively included in the database, it was found that HSS can successfully identify compounds with high reliability. The machine learning models and HSS algorithms are incorporated into our home-coded artificial intelligence screener for narcotic drugs and psychotropic substances standalone software that is equipped with a graphic user interface. The use of this software allows unknown controlled substances and NPSs to be identified in a convenient manner.

Diagnostic Framework for Automatic Classification and Visualization of Alzheimer’s Disease with Feature Extraction Using Wavelet Transform

Alzheimer’s Disease (AD) is a serious disease that destroys brain and is classified as the most widespread type of dementia. Manual evaluation of image scans relies on visual reading and semi-quantitative investigation of various human brain sections, leading to wrong diagnoses. Neuroimaging plays a significant part in AD detection, using image processing approaches that succeed the drawback of traditional diagnosis methods. Feature extraction is done through Wavelet Transform (WT). Feature selection is an important step in machine learning, where best features set from all possible features is determined. Mutual Information based feature selection (MI) and Correlation-based Feature Selection (CFS) captures the ‘correlation’ between random variables. Machine Learning techniques are broadly used in a classification problem, as it is simple, effective mechanisms and capability to train to contribute intelligence to the arrangement. Classifiers used in this proposed work are Artificial Neural Network (ANN), Random Forest, Convolutional Neural Network (CNN), and Wavelet-based CNN. The superior ability of ANN is high-speed processing achieved through extensive parallel implementation, and this has emphasized necessity of research in this field. CNN has encouraged tackling this issue. This work proves that wavelet-based CNN performs better with a classification accuracy of 91.87%, the sensitivity of 0.94 for normal brain and 0.88 for AD affected brain, the positive predictive value of 0.91 for normal brain and 0.92 for AD affected brain, and F measure of 0.92 for normal brain and 0.90 for AD affected brain on ADNI MRI dataset of the human brain in detecting AD.