Multilayer Perceptron Neural Network Research Articles

Exploring Quaternion Neural Network Loss Surfaces

This paper explores the superior performance of quaternion multi-layer perceptron (QMLP) neural networks over real-valued multi-layer perceptron (MLP) neural networks, a phenomenon that has been empirically observed but not thoroughly investigated. The study utilizes loss surface visualization and projection techniques to examine quaternion-based optimization loss surfaces for the first time. The primary contribution of this research is the statistical evidence that QMLP models yield smoother loss surfaces than real-valued neural networks, which are measured and compared using a robust quantitative measure of loss surface “goodness” based on estimates of surface curvature. Extensive computational testing validates the effectiveness of these surface curvature estimates. The paper presents a comprehensive comparison of the average surface curvature of a tuned QMLP model and a tuned real-valued MLP model on both a regression task and a classification task. The results provide strong support for the improved optimization performance observed in QMLPs across various problem domains.

Identification of cloud-to-ground lightning and intra-cloud lightning based on their radiated electric field signatures using different types of neural networks and machine learning classifiers

The scope of this paper is to test the efficiency of three different neural networks in the area of cloud-to-ground/intra-cloud lightning classification using a not very large data set of electric field data. Effective neural network learning verified for a small lightning event dataset but with whole 2-seconds particular record length is an important feature of fast and efficient procedures implemented into lightning location systems among which discrimination between cloud-to-ground and intra-cloud lightning is one of the most important task. Measurement data obtained at the Lightning Observatory in Rzeszow were used for the analysis. About 100 extremely low/medium frequency bandwidth electric field of cloud-to ground lightning registrations and the same number of intra-cloud lightning events were taken into account. During the study 81% of those registrations were dedicated for learning of neural networks while the remaining 19% were put into the testing set. All data were preselected manually with consideration of the electric field variation and information from the lightning location system database. Lightning events were selected uniformly with the well-pronounced presence of their basic components and reported distance to the registration station which ensured a wide variation of expected lightning electric field signatures. The multilayer perceptron neural network, the radial basis function neural network and the convolutional neural network were tested and compared in different configurations while distinguishing lightning. The results were compared with other conventional classification approaches, such as traditional machine learning methods as well as one a little more contemporary architecture, the long short-term memory neural network. The multilayer perceptron neural network achieved the best detection accuracy overall. Presented research is an extension of lightning identification studies available so far mainly based on the convolutional neural network by the results achieved using the multilayer perceptron neural network, the radial basis function neural network, the long short-term memory neural network and several machine learning-based classification methods.

Prediction of potentially toxic elements in water resources using MLP-NN, RBF-NN, and ANFIS: a comprehensive review.

Water resources are constantly threatened by pollution of potentially toxic elements (PTEs). In efforts to monitor and mitigate PTEs pollution in water resources, machine learning (ML) algorithms have been utilized to predict them. However, reviewstudies have not paid attention to the suitability of input variables utilized for PTE prediction. Therefore, the present review analyzed studies that employed three ML algorithms: MLP-NN (multilayer perceptron neural network), RBF-NN (radial basis function neural network), and ANFIS (adaptive neuro-fuzzy inference system) to predict PTEs in water. A total of 139 models were analyzed to ascertain the input variables utilized, the suitability of the input variables, the trends of the ML model applications, and the comparison of their performances. The present study identified seven groups of input variables commonly used to predict PTEs in water. Group 1 comprised of physical parameters (P), chemical parameters (C), and metals (M). Group 2 contains only P and C; Group 3 contains only P and M; Group 4 contains only C and M; Group 5 contains only P; Group 6 contains only C; and Group 7 contains only M. Studies that employed the three algorithms proved that Groups 1, 2, 3, 5, and 7 parameters are suitable input variables for forecasting PTEs in water. The parameters of Groups 4 and 6 also proved to be suitable for the MLP-NN algorithm. However, their suitability with respect to the RBF-NN and ANFIS algorithms could not be ascertained. The most commonly predicted PTEs using the MLP-NN algorithm were Fe, Zn, and As. For the RBF-NN algorithm, they were NO3, Zn, and Pb, and for the ANFIS, they were NO3, Fe, and Mn. Based on correlation and determinationcoefficients (R,R2), the overall order of performance of the three ML algorithms was ANFIS > RBF-NN > MLP-NN, even though MLP-NN was the most commonly used algorithm.

Analysis and Prediction of Urban Surface Transformation Based on Small Baseline Subset Interferometric Synthetic Aperture Radar and Sparrow Search Algorithm-Convolutional Neural Network-Long Short-Term Memory Model.

With the acceleration of urbanisation, urban areas are subject to the combined effects of the accumulation of various natural factors, such as changes in temperature leading to the thermal expansion or contraction of surface materials (rock, soil, etc.) and changes in precipitation and humidity leading to an increase in the self-weight of soil due to the infiltration of water along the cracks or pores in the ground. Therefore, the subsidence of urban areas has now become a serious geological disaster phenomenon. However, the use of traditional neural network prediction models has limitations when examining the causal relationships between time series surface deformation data and multiple influencing factors and when applying multiple influencing factors for predictive analyses. To this end, Sentinel-1A data from March 2017 to February 2023 were used as the data source in this paper, based on time series deformation data acquired using the small baseline subset interferometric synthetic aperture radar (SBAS-InSAR) technique. A sparrow search algorithm-convolutional neural network-long short-term memory (SSA-CNN-LSTM) neural network prediction model was built. The six factors of temperature, humidity, precipitation, and ground temperature at three different depths below the surface (5 cm, 10 cm, and 15 cm) were taken as the input of the model, and the surface deformation data were taken as the output of the neural network model. The correlation between the spatial and temporal evolution characteristics of the ground subsidence in urban areas and various influencing factors was analysed using grey correlation analysis, which proved that these six factors contribute to some extent to the deformation of the urban surface. The main urban area of Hohhot City, Inner Mongolia Autonomous Region, was used as the study area. In order to verify the efficacy of this neural network prediction model, the prediction effects of the multilayer perceptron (MLP), backpropagation (BP), and SSA-CNN-LSTM models were compared and analysed, with the values of the correlation coefficients of the feature points of A1, B1, and C1 being in the range of 0.92, 0.83, and 0.93, respectively. The results show that compared with the traditional MLP and BP neural network models, the SSA-CNN-LSTM model achieves a higher performance in predicting time series surface deformation data in urban areas, which provides new ideas and methods for this area of research.

Research on a Highway Passenger Volume Prediction Model Based on a Multilayer Perceptron Neural Network

Research on a Highway Passenger Volume Prediction Model Based on a Multilayer Perceptron Neural Network

Performance analysis of Machine Learning Algorithms to classify Software Requirements

Building a software system requires a clear understanding of its purpose and its operational characteristics. Functional requirements establish the system's purpose, while non-functional requirements define its operational performance aspects. It is essential to identify and classify requirements accurately to develop reliable software. In this research paper, we aim to classify functional and non-functional software requirements using different Machine Learning algorithms and techniques. We utilized four popular classification models, including Logistic Regression, Support Vector Machines, Decision Tree, and Random Forest Multi-layer Perceptron Neural Network to classify the requirements. To enhance the accuracy of the classification, we also applied a technique based on cosine similarity to verify if the custom string provided as input is related to software requirements. The addition of cosine similarity improved the accuracy of classification and reduced the misclassification of non-requirements.

Antimicrobial Activity Classification of Imidazolium Derivatives Predicted by Artificial Neural Networks

PurposeThis study assesses the Multilayer Perceptron (MLP) neural network, complemented by other Machine Learning techniques (CART, PCA), in predicting the antimicrobial activity of 140 newly designed imidazolium chlorides against Klebsiella pneumoniae before synthesis. Emphasis is on leveraging molecular properties for predictive analysis.MethodsClassification and regression decision trees (CART) identified the top 200 predictive molecular descriptors. Principal Component Analysis (PCA) reduced these descriptors to 5 components, retaining 99.57% of raw data information. Antimicrobial activity, categorized as high or low, was based on experimentally proven minimal inhibitory concentration (MIC), with a cut-point at MIC = 0.856 mol/L. A 12-fold cross-validation trained the MLP (architecture 5-12-2 with 5 Principal Components).ResultsThe MLP exhibited commendable performance, achieving almost 90% correct classifications across learning, validation, and test sets, outperforming models without PCA dimension reduction. Key metrics, including accuracy (0.907), sensitivity (0.905), specificity (0.909), and precision (0.891), were notably high. These results highlight the MLP model's efficacy with PCA as a high-quality classifier for determining antimicrobial activity.ConclusionsThe study concludes that the MLP neural network, along with CART and PCA, is a robust tool for predicting the antimicrobial activity class of imidazolium chlorides against Klebsiella pneumoniae. CART and PCA, used in this study, allowed input variable reduction without significant information loss. High classification accuracy and associated metrics affirm the method’s potential utility in pre-synthesis assessments, offering valuable insights for antimicrobial compound design.

Forecasting hotel cancellations through machine learning

AbstractAccurate and reliable forecasting of cancellations is important for successful revenue management in the tourism industry. The objective of this study is to develop classification models to predict hotel booking cancellations. The work involves a number of key steps, such as data preprocessing to properly prepare the data; feature engineering to identify relevant attributes to help improve the predictive ability of the models; hyperparameter settings of the models, including choice of optimizers and incorporation of dropout layers to avoid overfitting in the neural networks; potential overfitting is evaluated using K‐fold cross‐validation; and performance is analysed using the confusion matrix and various performance metrics. The algorithms used are Multilayer Perceptron Neural Network, Radial Basis Function Neural Network, Deep Neural Network, Decision Tree Classifier, Random Forest Classifier, Ada Boost Classifier and XgBoost Classifier. Finally, the results of all models are compared, visualizing Deep Neural Network and XgBoost as the most suitable models for predicting hotel reservation cancellations.

Carbon Emission Forecast Based on Multilayer Perceptron Network and STIRPAT Model

INTRODUCTION: It is of great research significance to explore whether China can achieve the "two-carbon target" on time. The MLP model combines nonlinear modeling principles with other techniques, possessing powerful adaptive learning capabilities, and providing a viable solution for carbon emission prediction.
 OBJECTIVES: This study models and forecasts carbon emissions in Jiangsu Province, one of China's largest industrial provinces, aiming to forecast whether Jiangsu province will achieve the two-carbon target on time plan and provide feasible pathways and theoretical foundations for achieving dual carbon goals.
 METHODS: Based on the analysis of the contributions of relevant indicators using the Grey Relational Analysis method, a comprehensive approach integrating the STIRPAT model, Logistic model, and ARIMA model is adopted. Ultimately, an MLP prediction model for carbon emission variations is established. Using this model, simulations are conducted to analyze the carbon emission levels in Jiangsu Province under different scenarios from 2021 to 2060.
 RESULTS: The time to reach carbon peak and the likelihood of achieving carbon neutrality vary under three scenarios. Under the natural scenario of no human intervention, achieving carbon neutrality is not feasible. While under human-made intervention scenarios including baseline and intervention scenarios, Jiangsu Province is projected to achieve the carbon neutrality target as scheduled, attaining the peak carbon goal, however, proves challenging to realize by the year 2030.
 CONCLUSION: The MLP model exhibits high accuracy in predicting carbon emissions. To expedite the realization of dual carbon goals, proactive government intervention is necessary.

Supported heterogeneous catalyst of the copper oxide nanoparticles and nanozeolite for binary dyes mixture degradation: Machine learning and experimental design

The present work aims to evaluate the photocatalytic activity of an alternative supported nanocatalyst (nANA@CuO-NPs) for a binary dye mixture (Crystal Violet and Methylene Blue, labeled CV:MB) under visible light and study to propose a model to predict the reaction pathway for the process. An experimental design (Central Composite Rotational Design with 3 factors and 2 levels – CCRD 23) was proposed to evaluate the ideal condition of the CV/MB photodegradation and four machine learning algorithms (Decision Tree, Random Forest, Xtreme Gradient Boosting and Artificial Neural Network-Multilayer Perceptron) were used to predict the main degradation products. The supported nanocatalyst was characterized by X-ray diffraction (XRD), Attenuated Total Reflectance-Fourier Transform Infrared (ATR-FTIR) spectroscopy, N2 porosimetry, Diffuse Reflectance Spectroscopy (DRS), Field Emission Gun – Scanning Electron Microscope (FEG-SEM), Scanning electron microscopy (SEM) with energy dispersive X-ray spectroscopy (EDX), zeta potential (ZP) and zero charge point (pHZCP). XRD diffractogram of the nANA@CuO-NPs showed the tenorite and analcime phases with crystallite size and microstrain ranging from 0.319 to 22.35 nm and 6.48–454 × 10−3. ATR-FTIR spectra indicated the presence of functional groups characteristic of copper oxide nanoparticles (CuO-NPs) and nanozeolite analcime (nANA) on the supported catalyst, confirming the impregnation method. SEM micrographs indicated a heterogeneous surface with clusters (CuO-NPs and nANA@CuO-NPs) and homogenous surface for nANA, while the EDX showed the nANA with Si/Al < 2 (low-silica nanozeolite) and nANA@CuO-NPs with the presence in the elemental composition of the catalytic support (nANA) and the photoactive phase (CuO-NPs). Supported nanocatalyst showed physio-chemical stability due to the electrostatic interactions (ZP = 51.5 ± 0.5 mV) and band gap energy (Eg) of the 1.67 eV allowing its applicability under visible radiation. XGB algorithm resulted in the best predictive model (R2 equals 0.8106 and 0.9880 for training and testing, RMSE < 6.0), confirming the obtention of carbon dioxide (m/z = 44), water (m/z = 18) and low-molar mass compounds at the final of the binary dye mixture degradation reaction. The ideal condition for the by CCRD 23 for the degradation of the dye mixture was [CV:MB] = 135 mg/L, [nANA@CuO-NPs] = 1 g/L and pH = 7 with 82.84 % (k = 0.0089 min−1). Therefore, the synthesized nanocatalyst proved its good photocatalytic activity toward a cationic dye mixture of CV and MB dyes. Furthermore, this work confirms the potentiality of machine learning algorithms to develop predictive models for the elucidation of the degradation reaction pathway of organic dyes.

Impacts of Feature Selection on Predicting Machine Failures by Machine Learning Algorithms

In the context of Industry 4.0, managing large amounts of data is essential to ensure informed decision-making in intelligent production environments. It enables, for example, predictive maintenance, which is essential for anticipating and identifying causes of failures in machines and equipment, optimizing processes, and promoting proactive management of human, financial, and material resources. However, generating accurate information for decision-making requires adopting suitable data preprocessing and analysis techniques. This study explores the identification of machine failures based on synthetic industrial data. Initially, we applied the feature selection techniques Principal Component Analysis (PCA), Minimum Redundancy Maximum Relevance (mRMR), Neighborhood Component Analysis (NCA), and Denoising Autoencoder (DAE) to the collected data and compared their results. In the sequence, a comparison among three widely known machine learning classifiers, namely Random Forest (RF), Support Vector Machine (SVM), and Multilayer Perceptron neural network (MLP), was conducted, with and without considering feature selection. The results showed that PCA and RF were superior to the other techniques, allowing the classification of failures with rates of 0.98, 0.97, and 0.98 for the accuracy, precision, and recall metrics, respectively. Thus, this work contributes by solving an industrial problem and detailing techniques to identify the most relevant variables and machine learning algorithms for predicting machine failures that negatively impact production planning. The findings provided by this study can assist industries in giving preference to employing sensors and collecting data that can contribute more effectively to machine failure predictions.

Prediction of Noise Levels According to Some Exploitation Parameters of an Agricultural Tractor: A Machine Learning Approach

The paper presents research on measuring and the possibility of prediction of noise levels on the left and right sides of the operator within the cabin of an agricultural tractor when moving across various agrotechnical surfaces, considering movement velocity and tire pressures while employing machine learning techniques. Noise level measurements were conducted on a LANDINI POWERFARM 100 type tractor, and aligned with standards (HRN ISO 5008, HRN ISO 6396 and HRN ISO 5131). The obtained noise values were divided into two data sets (left and right set) and processed using multiple linear regression (mlr) and three machine learning methods (gradient boosting machine (gbm); support vector machine using radial basis function kernel (svmRadial); monotone multi-layer perceptron neural network (monmlp)). The most accurate method, considering surfaces, from the left side data set—(R2 0.515–0.955); (RMSE 0.302–0.704); (MAE 0.225–0.488)—and the right side—(R2 0.555–0.955); (RMSE 0.180–0.969); (MAE 0.139–0.644)—was monmlp predominantly, and to a lesser extent svmRadial. On analyzing the total data sets from the left and right sides regarding surfaces, gbm emerged as the most accurate method. The application of machine learning methods demonstrated data accuracy, yet in future research, measurements on certain surfaces may need to be repeated multiple times potentially to improve accuracy further.

Methodology for Multi-Step Forecasting of Electricity Spot Prices Based on Neural Networks Applied to the Brazilian Energy Market

Forecasting electricity spot prices holds paramount significance for informed decision-making among energy market stakeholders. This study introduces a methodology utilizing a multilayer perceptron (MLP) neural network for multivariate electricity spot price prediction. The model underwent a feature selection process to identify the most influential predictors. In the validation phase, the model’s performance was evaluated using key metrics, including trend accuracy percentage index (TAPI), normalized root mean squared error (NRMSE), and mean absolute percentage error (MAPE). The results were obtained for a four-week forecast horizon in order to serve as an auxiliary tool to facilitate decision-making processes in the short-term energy market. The relevance of short-term electricity spot price forecasting lies in its direct impact on pricing strategies during energy contract negotiations, which allows for the making of assertive decisions in the energy trading landscape.

Investigating the rheological characteristics of alkali-activated concrete using contemporary artificial intelligence approaches

Abstract Using artificial intelligence-based tools, this research aims to establish a direct correlation between the alkali-activated concrete (AAC) mix design factors and their performances. More specifically, the machine learning system was fed new property data obtained from AAC mixes used in laboratory experiments. The rheological parameters (yield stress [static/dynamic] and plastic viscosity) of AAC were predicted using the multilayer perceptron neural network (MLPNN) and bagging ensemble (BE) models. In addition, the R 2 values, k-fold analyses, statistical checks, and the dissimilarity between the experimental and predicted compressive strength were employed to assess the performance of the created models. Also, the SHapley additive exPlanation (SHAP) approach was used for examining the relevance of influencing parameters. The BE approach was found to be significantly accurate in all prediction models, with R 2 greater than 0.90, and MLPNN models were found to be moderately precise, with R 2 slightly below 0.90. However, the error assessment through statistical checks and k-fold analysis also validated the higher precision of BE models over the MLPNN models. Building models that can calculate rheological properties of AAC for different values of input parameters could save a lot of time and money compared to doing the tests in a laboratory. In order to ascertain the required amounts of raw materials of AAC, investigators, as well as businesses, may find the SHAP study helpful.

Modeling the decay of free residual chlorine in water distribution networks in Brazilian rural communities using artificial neural network

Disinfecting water with chlorine is crucial for maintaining sanitation and preventing waterborne diseases. However, managing Free Residual Chlorine (FRC) concentrations becomes challenging in water supply systems, as its levels can decrease throughout the distribution network, impacting water quality and public health. By leveraging water quality data from the Water Quality Surveillance Information System for Human Consumption (Sisagua) and employing the variable selection method t-Distributed Stochastic Neighbor Embedding, we crafted a Multilayer Perceptron (MLP) neural network. This MLP, fine-tuned with weighted mean square error (WMSE), was designed to evaluate the decline of free residual chlorine in water distribution networks within rural Brazilian communities. Training the model with the weighted mean squared error (WMSE) increased the average hit rate from 3.6 % to 5.0 %. In contrast to the traditional colorimetric method for predicting free residual chlorine, the proposed methodology showed superior performance, with accuracy rates of 85.45 %, 95.58 % and 99.84 % for equipment accuracies of 0.5 mg. /L, 1.0 mg/L and 2.0 mg/L, respectively. The MLP approach not only anticipates nonconformities related to FRC but also aids in adjusting chlorine dosage proactively.

Geometrically-informed predictive modeling of melt pool depth in laser powder bed fusion using deep MLP-CNN and metadata integration

The laser powder bed fusion (LPBF) process has the ability to manufacture intricate geometries. However, the fabrication of complex geometries using the LPBF process for industrial applications remains challenging in terms of achieving consistent microstructure and mechanical properties. Experimental investigations show that in addition to process parameters, geometrical factors such as shape and size of parts impact the thermal history, microstructure, defect structure, and thus mechanical and fatigue properties. As a result, the standardization, qualification, and certification of additively manufactured (AM-ed) parts encounters substantial obstacles posed by the obscure impact of design parameters on component quality. In this paper, an artificial intelligence framework is proposed to predict the melt pool depth of the additively manufactured parts considering the geometrical information of the parts. More specifically, a deep Multilayer Perceptron and Convolutional Neural Network (MLP-CNN) framework is designed to predict melt pool depth for any part geometries when printing metallic parts via LPBF using metadata (i.e., numerical and image data). Quite often the melt pool depth generated during the AM processes is employed as a surrogate for thermal, defect- and micro-structural signatures. In melt pool depth prediction of complex geometries in LPBF, a digital twin environment using a finite element model (FEM) is developed and validated using experimental melt pool measurements for different geometries within multi-tracks and layers. The FEM was in agreement with experimentation with an error of less than 15%. Through the process simulation, a large dataset of melt pool depths was obtained for various part geometries and different process parameters. Next, the MLP-CNN framework was established to identify the parallel impact of part design and process parameters on melt pool depths. To enhance the performance and robustness of this model, data augmentation has been implemented by rotating and transferring geometries to artificially expand the dataset. Data augmentation helps to mitigate overfitting and promote better generalization, especially in the context of limited training data. After training, the proposed model is found to give accurate melt pool prediction even for new geometries not considered during training. The fusion of CNN and MLP has led the melt pool predictions for unseen geometries with an accuracy of 95%.

Data‐Driven Cz–Si Scale‐Up under Conditions of Partial Similarity

AbstractIn Cz–Si growth, the shape of the solid–liquid interface and the v/G ratio significantly impact crystal quality. This study utilizes a data‐driven approach, employing multilayer perceptron (MLP) neural networks and Bayesian optimization, to investigate the scale‐up process of Cz–Si under conditions of partial similarity. The focus is on exploring the influence of various process and furnace geometry parameters, as well as radiation shield material properties, on the critical measures of crystal quality. Axisymmetric CFD modeling produces 340 sets of 18D raw data, from which 14‐dimensionless derived data tuples are generated for the design and training of the MLP. The best MLP obtained demonstrates the ability to accurately assess the complex nonlinear dependencies among dimensionless numbers derived from CFD data and, on the output side, interface deflection and v/G. These relationships, crucial for scale‐up, are successfully generalized across a wide range of parameters.

The prediction of precipitation changes in the Aji-Chay watershed using CMIP6 models and the wavelet neural network

ABSTRACT Greenhouse gases affect climate system disturbances. This research employs sixth generation CMIP6 models in the SSP5.85 scenario and extends the use of the neural wavelet network to predict precipitation variations for the future (2025–2065). Kendall's trend test is used to assess changes in precipitation trends for observed and projected periods. An analysis of variance (ANOVA) validates models under SSP5.85 by comparing observed precipitation with model predictions. A multi-layer perceptron neural network assesses climate change's impact on future precipitation. Findings indicate future precipitation is projected to fluctuate from −0.146 to over −2.127 mm compared to the baseline period. The observed period showed a significant 3.37% monthly precipitation decrease within the watershed. The CanESM5 model predicts a 3.916 reduction in precipitation with 95% confidence, while INM-CM4-8 and MRI-ESM2-0 models are less certain. The minor difference between CanESM5's predicted (−5.91) and observed (−5.05) precipitation suggests a slight variance. On the other hand, the wavelet neural network (WNN) model predicts that precipitation in this region will increase in the future. In general, this study predicts a decrease in precipitation for the Aji-Chay watershed in Iran over the next decade, could lead to serious issues like lower crop yields, rising food prices, and even droughts.

Automated real-time prediction of geological formation tops during drilling operations: an applied machine learning solution for the Norwegian Continental Shelf

Accurate prediction of geological formation tops is a crucial task for optimizing hydrocarbon exploration and production activities. This research investigates and conducts a comprehensive comparative analysis of several advanced machine learning approaches tailored for the critical application of geological formation top prediction within the complex Norwegian Continental Shelf (NCS) region. The study evaluates and benchmarks the performance of four prominent machine learning models: Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Random Forest ensemble method, and Multi-Layer Perceptron (MLP) neural network. To facilitate a rigorous assessment, the models are extensively evaluated across two distinct datasets - a dedicated test dataset and a blind dataset independent for validation. The evaluation criteria revolve around quantifying the models' predictive accuracy in successfully classifying multiple geological formation top types. Additionally, the study employs the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm as a baseline benchmarking technique to contextualize the relative performance of the machine learning models against a conventional clustering approach. Leveraging two model-agnostic feature importance analysis techniques - Permutation Feature Importance (PFI) and Shapley Additive exPlanations (SHAP), the investigation identifies and ranks the most influential input variables driving the predictive capabilities of the models. The comprehensive analysis unveils the MLP neural network model as the top-performing approach, achieving remarkable predictive accuracy with a perfect score of 0.99 on the blind validation dataset, surpassing the other machine learning techniques as well as the DBSCAN benchmark. However, the SVM model attains superior performance on the initial test dataset, with an accuracy of 0.99. Intriguingly, the PFI and SHAP analyses converge in consistently pinpointing depth (DEPT), revolution per minute (RPM), and Hook-load (HKLD) as the three most impactful parameters influencing model predictions across the different algorithms. These findings underscore the potential of sophisticated machine learning methodologies, particularly neural network-based models, to significantly enhance the accuracy of geological formation top prediction within the geologically complex NCS region. However, the study emphasizes the necessity for further extensive testing on larger datasets to validate the generalizability of the high performance observed. Overall, this research delivers an exhaustive comparative evaluation of state-of-the-art machine learning techniques, offering critical insights to guide the optimal selection, development, and real-world deployment of accurate and reliable predictive modeling strategies tailored for hydrocarbon exploration and reservoir characterization endeavors in the NCS.Graphical abstract

Artificial Neural Network for Automated Keratoconus Detection Using a Combined Placido Disc and Anterior Segment Optical Coherence Tomography Topographer.

To assess the efficacy of an automated program for keratoconus and keratoconus suspect detection based on corneal measurements provided by a combined Placido disc and anterior segment optical coherence tomography (OCT) topographer. In a multicentric cross-sectional study, an artificial neural network (ANN) was created using 6677 eyes from an equal number of patients (classified as 2663 normal eyes, 1616 keratoconus eyes, 210 keratoconus suspect eyes, 1519 myopic postoperative eyes, and 669 abnormal eyes). Each group was randomly divided into a training set (70% of the dataset) and a validation set (the remaining 30%). A multilayer perceptron network with a backpropagation learning algorithm was developed for the study. Indexes used to train the ANN were based on curvature and elevation of both the anterior and posterior corneal surfaces and the new corneal OCT indexes-based on corneal, stromal, and epithelial thicknesses. For keratoconus detection, our ANN showed an accuracy of 98.6%, precision of 96%, recall of 97.9%, and F1-score of 96.9%. For keratoconus suspect detection, our ANN showed an accuracy of 98.5%, precision of 83.6%, recall of 69.7%, and F1-score of 76%. Compared to previous literature, the addition of new OCT-based epithelial and stromal thickness indexes improves ANN detection capacity of keratoconus suspect eyes. For already stablished keratoconus our ANN detection capacity is excellent, but equivalent to previous evidence without incorporating such new OCT-based indexes. OCT-based epithelial and stromal thickness indexes improve ANN detection capacity of keratoconus on its early stages.