Multilayer Perceptron Research Articles

How Resilient Are Kolmogorov–Arnold Networks in Classification Tasks? A Robustness Investigation

Kolmogorov–Arnold Networks (KANs) are a novel class of neural network architectures based on the Kolmogorov–Arnold representation theorem, which has demonstrated potential advantages in accuracy and interpretability over Multilayer Perceptron (MLP) models. This paper comprehensively evaluates the robustness of various KAN architectures—including KAN, KAN-Mixer, KANConv_KAN, and KANConv_MLP—against adversarial attacks, which constitute a critical aspect that has been underexplored in current research. We compare these models with MLP-based architectures such as MLP, MLP-Mixer, and ConvNet_MLP across three traffic sign classification datasets: GTSRB, BTSD, and CTSD. The models were subjected to various adversarial attacks (FGSM, PGD, CW, and BIM) with varying perturbation levels and were trained under different strategies, including standard training, adversarial training, and Randomized Smoothing. Our experimental results demonstrate that KAN-based models, particularly the KAN-Mixer, exhibit superior robustness to adversarial attacks compared to their MLP counterparts. Specifically, the KAN-Mixer consistently achieved lower Success Attack Rates (SARs) and Degrees of Change (DoCs) across most attack types and datasets while maintaining high accuracy on clean data. For instance, under FGSM attacks with ϵ=0.01, the KAN-Mixer outperformed the MLP-Mixer by maintaining higher accuracy and lower SARs. Adversarial training and Randomized Smoothing further enhanced the robustness of KAN-based models, with t-SNE visualizations revealing more stable latent space representations under adversarial perturbations. These findings underscore the potential of KAN architectures to improve neural network security and reliability in adversarial settings.

Artificial Intelligence in Audiology: A Scoping Review of Current Applications and Future Directions

The integration of artificial intelligence (AI) into medical disciplines is rapidly transforming healthcare delivery, with audiology being no exception. By synthesizing the existing literature, this review seeks to inform clinicians, researchers, and policymakers about the potential and challenges of integrating AI into audiological practice. The PubMed, Cochrane, and Google Scholar databases were searched for articles published in English from 1990 to 2024 with the following query: “(audiology) AND (“artificial intelligence” OR “machine learning” OR “deep learning”)”. The PRISMA extension for scoping reviews (PRISMA-ScR) was followed. The database research yielded 1359 results, and the selection process led to the inclusion of 104 manuscripts. The integration of AI in audiology has evolved significantly over the succeeding decades, with 87.5% of manuscripts published in the last 4 years. Most types of AI were consistently used for specific purposes, such as logistic regression and other statistical machine learning tools (e.g., support vector machine, multilayer perceptron, random forest, deep belief network, decision tree, k-nearest neighbor, or LASSO) for automated audiometry and clinical predictions; convolutional neural networks for radiological image analysis; and large language models for automatic generation of diagnostic reports. Despite the advances in AI technologies, different ethical and professional challenges are still present, underscoring the need for larger, more diverse data collection and bioethics studies in the field of audiology.

An Artificial Intelligence Approach to Predict Physical Properties of Liquid Hydrocarbons

Accurate physical property prediction of newly developed compounds is vital across various industrial sectors, particularly for the customization of fuels and additives. Artificial intelligence (AI) has recently emerged as a best practice in numerous industrial fields because of its capacity for swift and precise calculations. While conventional methods such as group contribution models have been used to estimate physical properties from molecular structure, AI offers significant potential for improving the predictive accuracy. Thus, this work focuses on developing an AI model to predict key properties – boiling points, melting points, and flashpoints – of various hydrocarbons, to demonstrate the AI's superior predictive capabilities. A dataset consisting of 202 organic compounds was created and multilayer perceptron (MLP) neural networks were employed to estimate these properties using atomic numbers, functional groups, and molecular complexity as inputs. The model's performance was evaluated and compared against conventional group contribution methods on the same dataset. The AI model was further tested on new acetal compounds, revealing its broader applicability in both fuel and chemical sectors. Results show that the AI outperformed conventional methods, excelling in 5 out of 8 hydrocarbon types for boiling points, 7 for melting points, and all 8 for flashpoints.

Optimizing pulmonary chest x-ray classification with stacked feature ensemble and swin transformer integration

This research presents an integrated framework designed to automate the classification of pulmonary chest x-ray images. Leveraging convolutional neural networks (CNNs) with a focus on transformer architectures, the aim is to improve both the accuracy and efficiency of pulmonary chest x-ray image analysis. A central aspect of this approach involves utilizing pre-trained networks such as VGG16, ResNet50, and MobileNetV2 to create a feature ensemble. A notable innovation is the adoption of a stacked ensemble technique, which combines outputs from multiple pre-trained models to generate a comprehensive feature representation. In the feature ensemble approach, each image undergoes individual processing through the three pre-trained networks, and pooled images are extracted just before the flatten layer of each model. Consequently, three pooled images in 2D grayscale format are obtained for each original image. These pooled images serve as samples for creating 3D images resembling RGB images through stacking, intended for classifier input in subsequent analysis stages. By incorporating stacked pooling layers to facilitate feature ensemble, a broader range of features is utilized while effectively managing complexities associated with processing the augmented feature pool. Moreover, the study incorporates the Swin Transformer architecture, known for effectively capturing both local and global features. The Swin Transformer architecture is further optimized using the artificial hummingbird algorithm (AHA). By fine-tuning hyperparameters such as patch size, multi-layer perceptron (MLP) ratio, and channel numbers, the AHA optimization technique aims to maximize classification accuracy. The proposed integrated framework, featuring the AHA-optimized Swin Transformer classifier utilizing stacked features, is evaluated using three diverse chest x-ray datasets-VinDr-CXR, PediCXR, and MIMIC-CXR. The observed accuracies of 98.874%, 98.528%, and 98.958% respectively, underscore the robustness and generalizability of the developed model across various clinical scenarios and imaging conditions.

Unsupervised Fault Detection in a Controlled Conical Tank

Current trends in the Industrial Internet of Things (IIoT) have increased the sensorization of systems, thus increasing data availability to apply data-driven fault detection and diagnosis techniques to monitor these systems. In this work, we show the capabilities of an information-driven method for detecting and quantifying faults in a subsystem common among a broad range of industries: the conical tank. Our main experiment consists of using a simple black-box model (multi-layer perceptron -- MLP) to capture the dynamics of a PID-controlled conical tank built in Simulink and then induce pump failures of different severities; the derived data-driven indicators that we developed increase with the severity of the fault validating its usefulness in this controlled setting. A complementary experiment is carried out to enrich our analysis; this consists of simulating an open-loop discrete-time version of the conical tank to explore a range of fault severity and analyze the distribution of the indicators across this range. All our results show the applicability of the data-driven fault monitoring method in conical tanks subjected to either open- or closed-loop operation.

Interpretability as Approximation: Understanding Black-Box Models by Decision Boundary

Currently, interpretability methods focus more on less objective human-understandable semantics. To objectify and standardize interpretability research, in this study, we provide notions of interpretability based on approximation theory. We first define explainable models in terms of explicitness and then use completeness to define interpretability, thereby turning interpretability into the process of approximating black-box models with interpretable models. In particular, we think that the decision boundary of a classification model is equivalent to its interpretability. Next, we implement this approximation interpretation on multilayer perceptrons (MLPs) and then propose to use the MLP as a universal interpreter to explain other complex black-box models. Compared to the LIME method, which can only extract local linear features, our method is global and therefore termed as GIME. Extensive experiments demonstrate the effectiveness of our approaches.

Thermal Runaway Warning of Lithium Battery Based on Electronic Nose and Machine Learning Algorithms

Characteristic gas detection can be an efficient way to predict the degree of thermal runaway of a lithium battery. In this work, a sensor array consisting of three commercial MOS sensors was employed to discriminate between three target gases, CO, H2 and a mixture of the two, which are characteristic gases released during the thermal runaway of lithium batteries. In this work, an integrated model that makes the classification stage results one of the feature inputs for the concentration regression stage was employed, successfully reducing the RMSE of the concentration regression results. In addition, we also explored the influence of the selection of the response time length on the classification and regression tasks, achieving the best results in a short time through the optimum algorithm. To assess the impact of time duration sensor data on the results, we selected four time windows of different length and extracted the corresponding sensor response data for subsequent processing. Initially, principal component analysis (PCA) was used to visualise the clustering of the three target gas samples at room temperature, providing a preliminary data analysis. For the classification phase, we chose three classification algorithms—MLP (Multilayer Perceptron), ELM (Extreme Learning Machine), and SVM (Support Vector Machine)—and performed a comprehensive comparison of their classification and generalisation abilities using grid search for hyperparameter optimisation and five-fold cross-validation. The results demonstrated that MLP achieved 99.23% classification accuracy during the 20 s response period. In the concentration regression phase, we combined the classification results with the raw features to create a new feature set, which was then input into a multi-output MLP regression model. The root mean square error (RMSE) employing the new feature set was used to measure the prediction error. Ultimately, the findings showed that the input of combined features significantly reduced the regression error for the mixed gas.

Optimizing anemia management using artificial intelligence for patients undergoing hemodialysis

Patients with end-stage kidney disease (ESKD) frequently experience anemia, and maintaining hemoglobin (Hb) levels within a targeted range using erythropoiesis-stimulating agents (ESAs) is challenging. This study introduces a gated recurrent unit-attention-based module (GAM) for efficient anemia management among patients undergoing chronic dialysis and proposes a novel alert system for anticipating the need for red blood cell transfusions. Data on demographic characteristics, dialysis metrics, drug administration, laboratory tests, and transfusion history were retrospectively collected from patients undergoing hemodialysis at Kangwon National University Hospital between 2017 and 2022. After preprocessing, a final dataset of 252 patients was used for model training. Our model functions in two major phases: (1) Hb level prediction and ESA dose recommendation and (2) transfusion alert framework. The GAM model outperformed traditional machine learning algorithms, including linear regression, XGBoost, and multilayer perceptron, in predicting Hb levels (R-squared value = 0.60). The model also demonstrated a recommendation accuracy of 0.78 compared to that of clinical experts, indicating a high degree of concordance with the ESA dosing recommendations. Additionally, the model exhibited considerably high accuracy (0.99) for transfusion alarms. Thus, the GAM model holds promise for improving anemia management in patients with ESKD by optimizing ESA dosages and providing timely transfusion alerts.

Using DL Models in the Service Layer to Enhance the Fault Tolerance of IoT Networks

In an IoT network, the networked servers form a service layer, providing services to the users and the devices. The request to the service servers is routed through the gateway on one side of the services layer and the networked controllers on the other side. Data are transported from the sensors/devices through cluster heads en route to base stations and the controllers to the service servers, where the data are processed and sent for storage in the cloud through gateways. When any device is broken down or becomes non-operational, the inputs are not sensed, creating a gap in the data. The data transmitted from the devices would then become an incomplete flow; such data are not suitable for undertaking data analytics or predictions. The missing data must be first identified as the data flow and estimated or predicated to complete the data before they are transmitted through the cloud for storage and subsequent retrievals. This paper proposes a recurrent (RNN) neural network to predict the missing data. Two models are tested to predict the missing data: the multi-layer perceptron (MLP) model and a long short-term memory (LSTM)-based RNN model. The RNN-based model provides 99.66% accurate data prediction compared to other models.

Evolving Transparent Credit Risk Models: A Symbolic Regression Approach Using Genetic Programming

Credit scoring is a cornerstone of financial risk management, enabling financial institutions to assess the likelihood of loan default. However, widely recognized contemporary credit risk metrics, like FICO (Fair Isaac Corporation) or Vantage scores, remain proprietary and inaccessible to the public. This study aims to devise an alternative credit scoring metric that mirrors the FICO score, using an extensive dataset from Lending Club. The challenge lies in the limited available insights into both the precise analytical formula and the comprehensive suite of credit-specific attributes integral to the FICO score’s calculation. Our proposed metric leverages basic information provided by potential borrowers, eliminating the need for extensive historical credit data. We aim to articulate this credit risk metric in a closed analytical form with variable complexity. To achieve this, we employ a symbolic regression method anchored in genetic programming (GP). Here, the Occam’s razor principle guides evolutionary bias toward simpler, more interpretable models. To ascertain our method’s efficacy, we juxtapose the approximation capabilities of GP-based symbolic regression with established machine learning regression models, such as Gaussian Support Vector Machines (GSVMs), Multilayer Perceptrons (MLPs), Regression Trees, and Radial Basis Function Networks (RBFNs). Our experiments indicate that GP-based symbolic regression offers accuracy comparable to these benchmark methodologies. Moreover, the resultant analytical model offers invaluable insights into credit risk evaluation mechanisms, enabling stakeholders to make informed credit risk assessments. This study contributes to the growing demand for transparent machine learning models by demonstrating the value of interpretable, data-driven credit scoring models.

Leveraging convolutional neural networks and hashing techniques for the secure classification of monkeypox disease

The World Health Organization declared a state of emergency in 2022 because of monkeypox. This disease has raised international concern as it has spread beyond Africa, where it is endemic. The global community has shown attention and solidarity in combating this disease as its daily increase becomes evident. Various skin symptoms appear in people infected with this disease, which can spread easily, especially in a polluted environment. It is difficult to diagnose monkeypox in its early stages because of its similarity with the symptoms of other diseases such as chicken pox and measles. Recently, computer-aided classification methods such as deep learning and machine learning within artificial intelligence have been employed to detect various diseases, including COVID-19, tumor cells, and Monkeypox, in a short period and with high accuracy. In this study, we propose the CanDark model, an end-to-end deep-learning model that incorporates cancelable biometrics for diagnosing Monkeypox. CanDark stands for cancelable DarkNet-53, which means that DarkNet-53 CNN is utilized for extracting deep features from Monkeypox skin images. Then a cancelable method is applied to these features to protect patient information. Various cancelable techniques have been evaluated, such as bio-hashing, multilayer perceptron (MLP) hashing, index-of-maximum Gaussian random projection-based hashing (IoM-GRP), and index-of-maximum uniformly random permutation-based hashing (IoM-URP). The proposed approach’s performance is evaluated using various assessment issues such as accuracy, specificity, precision, recall, and fscore. Using the IoM-URP, the CanDark model is superior to other state-of-the-art Monkeypox diagnostic techniques. The proposed framework achieved an accuracy of 98.81%, a specificity of 98.73%, a precision of 98.9%, a recall of 97.02%, and fscore of 97.95%.

A Machine Learning Framework for Classroom EEG Recording Classification: Unveiling Learning-Style Patterns

Classroom EEG recordings classification has the capacity to significantly enhance comprehension and learning by revealing complex neural patterns linked to various cognitive processes. Electroencephalography (EEG) in academic settings allows researchers to study brain activity while students are in class, revealing learning preferences. The purpose of this study was to develop a machine learning framework to automatically classify different learning-style EEG patterns in real classroom environments. Method: In this study, a set of EEG features was investigated, including statistical features, fractal dimension, higher-order spectra, entropy, and a combination of all sets. Three different machine learning classifiers, random forest (RF), K-nearest neighbor (KNN), and multilayer perceptron (MLP), were used to evaluate the performance. The proposed framework was evaluated on the real classroom EEG dataset, involving EEG recordings featuring different teaching blocks: reading, discussion, lecture, and video. Results: The findings revealed that statistical features are the most sensitive feature metric in distinguishing learning patterns from EEG. The statistical features and RF classifier method tested in this study achieved an overall best average accuracy of 78.45% when estimated by fivefold cross-validation. Conclusions: Our results suggest that EEG time domain statistics have a substantial role and are more reliable for internal state classification. This study might be used to highlight the importance of using EEG signals in the education context, opening the path for educational automation research and development.

Multi-Step Ahead Water Level Forecasting Using Deep Neural Networks

Stream gauge height (water level) is a significant indicator for forecasting future floods. Flooding occurs when the water level exceeds the flood stage. Predicting imminent floods can save lives, protect infrastructure, and improve road traffic management and transportation. Deep neural networks have been increasingly used in this domain due to their predictive capabilities in capturing complex features and interdependencies. This study employs four distinct models—Multi-Layer Perceptron (MLP), Long Short-Term Memory (LSTM), transformer, and LSTNet—with MLP serving as the baseline model to forecast water levels. The models are trained using data from 20 distinct river gages across the state of Missouri to ensure consistent performance. Random search optimization is employed for hyperparameter tuning. The prediction intervals are set at 4, 6, 8, and 10 (each interval equivalent to 30 min) to ensure that performance results are robust and not due to random weight initialization or suboptimal hyperparameters and are consistent throughout different prediction intervals. The findings of this study indicate that the LSTNet model leads to a better performance than the other models, with a median RMSE of 0.00724, 0.00959, 0.01204, and 0.01230 for the 4, 6, 8, and 10 intervals, respectively. As climate change leads to localized storms driven by atmospheric shifts, water level fluctuations are becoming increasingly extreme, further exacerbating data drift in real-world datasets. The LSTNet model demonstrates superior performance in terms of RMSE, MAE, and the correlation coefficient across all prediction intervals when forecasting water levels under data drift conditions.

Real-life boxing activity recognition with smartphones using attention assisted deep learning models

Mobile computing technologies play a pivotal role in narrowing the gap between athletes and sports monitoring systems, while also advancing towards intelligent and automatic supervision of individual sports activities. The widespread presence of smartphones equipped with multifunctional sensors has enabled the acquisition and analysis of data, facilitating Human Activity Recognition (HAR). Although recognising sports activities using a smart mobile phone’s accelerometer is widely used, traditional methods struggle with intricate and real time activities due to the high-dimensional sensor data. This study introduces a smartphone-based architecture for HAR utilising inertial accelerometers to record athlete’s game actions data sequences, extract relevant features and derive the player’s activity data through multiple three-axis accelerometers. In this paper an Attention assisted Deep Convolution Neural Network based Bi-LSTM (AT-DCNN-BL) model is proposed. The raw data undergoes sliding window pre-processing technique, the deep learning framework of deep convolution neural network layers helps in automatic feature extraction, the bi-LSTM structure process the boxing activity data and the attention is focussed over the boxing activity classification data. The attention mechanism redistributes the weights of the extracted features which aids in increasing the classification accuracy. Other classification models like Deep Convolutional Neural Network (DCNN), Bi-directional Long Short-Term Memory (bi-LSTM), Multi-Layer Perceptron Neural Network (MLPNN) and Random Forest (RF) are also applied to the dataset collected from a mobile sensor. The study explores training of deep learning methods and illustrates the superiority of the proposed approach over others using the collected mobile sensor dataset. The AT-DCNN-BL model achieved a higher classification accuracy compared to the other models. The AT-DCNN-BL model outperformed all the other models achieving a higher accuracy rate (92.71%).

Low-resource dynamic loading identification of nonlinear system using pretraining

To address the challenge of load identification in nonlinear systems, an audio neural network-based method called WaveNet is proposed that leverages its capability to capture long-term dependencies in mechanical systems, enabling accurate load identification. Unlike traditional dynamic load identification methods that often encounter difficulties with matrix solutions, this approach takes advantage of WaveNet’s capabilities, enhancing both accuracy and efficiency. We integrate pre-training and transfer learning techniques to address the data scarcity challenges often encountered in real-world engineering applications. By transferring features across distributed datasets, this method reduces the dependency on single-task data, thereby improving model robustness. The performance of the WaveNet method is rigorously evaluated against traditional benchmarks such as Multilayer Perceptron (MLP) and Convolutional Neural Network (CNN) benchmarks under random load conditions applied to a complex structural framework. The proposed method achieves a root mean squared error (RMSE) of 1.521 and a determination coefficient (R²) of 0.996 in the random load case, demonstrating superior accuracy compared to other approaches. Moreover, its applicability is verified through simulations of both impact load and harmonic load scenarios, showcasing the effectiveness of transfer learning in overcoming domain discrepancies. Finally, the method is tested in random experiments to validate its engineering applicability. The results highlight the significant accuracy improvement in low-resource tasks achieved through pre-training, showcasing the potential and value of the proposed method.

MODEL OF MAXIMAL WEIGHTS INVERSE CHAINS FOR THE ANALYSIS OF THE INFLUENCE FACTORS OF THE SOFTWARE COMPLEXES SUPPORT

Context. The problem of identification, formation and restoration of the boundaries of influencing factors, lost as a result of the implementation of multi-layer perceptron models into the models of subjective perception of the object of software complexes support, as well as the applied practical problem of primary monitoring of the frequency manifestation of a given influencing factor in the post-real-time mode, is considered. The object of research is the influencing factors of support of software complexes. Objective – the goal of the work is to develop a model of inverse chains of maximum weights for the analysis of influencing factors of the software complexes support. Method. A model of maximal weights inverse chains for the analysis of the influence factors of the software complexes support was developed for the analysis of the influencing factors of the software complexes support. The developed model provides possibility to identify and form feedback chains of maximum weights for the identification and further analysis of influencing factors that are reflected into the results of the object perception (the supported software complex or its support processes), by the relevant subjects of interaction which directly or indirectly interact with it. Results. Results of the resolved applied practical problem of primary monitoring of the frequency manifestation of a given influencing factor in the post-real-time mode have been provided as an example of the applied practical use of the developed model. The output results of the developed models functioning – are the reverse chains of maximum weights. In the future, the results obtained by the developed model are used to solve the applied-scientific problem of identification, formation and restoration of the boundaries of influencing factors, lost as a result of the implementation of the appropriate models of multilayer perceptron inside the models of subjective perception of the software complexes support. So the developed model of maximal weights inverse chains for the analysis of the influence factors of the software complexes support resolves this applied-scientific problem, initially caused by the implementation of the corresponding multilayer perceptron models inside the model of the subjective perception of the object of software complexessupport. The developed model provides the possibility of carrying out a qualitative analysis of the transformation of the input characteristics of the object of support into the output resulting characteristics of its subjective perception. Conclusions. Developed model allows to resolve the described problems. At the same time, the developed model improves the classical understanding of multilayer perceptron artificial neural networks, as it introduces an additional value to the neurons of hidden layers, which (starting from now) are able to perform a completely new role of influencing factors markers, while in the classical understanding of multilayer perceptron artificial neural networks they did not perform any functions other than arithmetic to ensure the possibility of correct learning and functioning of a multilayer perceptron artificial neural networks.

Sediment Simulation Using Multi Layer Perceptron (MLP), Co-Active Neuro-Fuzzy Inference System (CANFIS) and Multiple Linear Regression Tecniques (MLR) for Hurdag Watershed

Sediment modeling plays a crucial role in sustainable water resources planning, development, and management. Techniques like the Multilayer Perceptron (MLP), Co-active Neuro-Fuzzy Inference System (CANFIS), and Multiple Linear Regression (MLR) have proven effective for sediment modeling and forecasting. This study aimed to develop and assess the applicability of MLP, CANFIS, and MLR models by training and testing them during the monsoon season (June to September) for the Hurdag watershed in the Damodar-Barakar basin, located in Hazaribagh district, Jharkhand, India. Daily rainfall, runoff (or streamflow), and sediment concentration data from 1997 to 2006 were used, with the data split into two sets: training set (1997–2004) and a testing set (2005–2006). The analysis was conducted using NeuroSolution 5.0 software and Microsoft Excel for performance evaluation indices. The best input combinations for sediment yield simulation were identified, and 10 optimal models were selected from 31 different input combinations. These input combinations were applied to the network for training using the back propagation algorithm for MLP and Gaussian and generalized bell membership functions for CANFIS models. Multiple networks were trained individually, with the most accurate predictions during testing being chosen as the best models. The models’ performance was evaluated using statistical indices such as root mean squared error (RMSE), coefficient of efficiency (CE), and correlation coefficient (r). The results showed that MLP and CANFIS models performed best in predicting sediment concentration for the Hurdag watershed, whereas MLR models showed poor performance for the given dataset. Specifically, sediment concentration for the current day could be modeled using current day rainfall and runoff data (MLP-7), while runoff could be simulated using the previous day's rainfall data (MLP-2).

Securing transactions: a hybrid dependable ensemble machine learning model using IHT-LR and grid search

Financial institutions and businesses face an ongoing challenge from fraudulent transactions, prompting the need for effective detection methods. Detecting credit card fraud is crucial for identifying and preventing unauthorized transactions. While credit card fraud incidents are relatively rare, they can result in substantial financial losses, particularly due to the high monetary value associated with fraudulent transactions. Timely detection of fraud enables investigators to take swift actions to mitigate further losses. However, the investigation process is often time-consuming, limiting the number of alerts that can be thoroughly examined each day. Therefore, the primary objective of a fraud detection model is to provide accurate alerts while minimizing false alarms and missed fraud cases. In this paper, we introduce a state-of-the-art hybrid ensemble (ENS) dependable machine learning (ML) model that intelligently combines multiple algorithms with proper weighted optimization using grid search, including decision tree (DT), random forest (RF), K-nearest neighbor (KNN), and multilayer perceptron (MLP), to enhance fraud identification. To address the data imbalance issue, we employ the instant hardness threshold (IHT) technique in conjunction with logistic regression (LR), surpassing conventional approaches. Our experiments are conducted on a publicly available credit card dataset comprising 284,807 transactions. The proposed model achieves impressive accuracy rates of 99.66%, 99.73%, 98.56%, and 99.79%, and a perfect 100% for the DT, RF, KNN, MLP and ENS models, respectively. The hybrid ensemble model outperforms existing works, establishing a new benchmark for detecting fraudulent transactions in high-frequency scenarios. The results highlight the effectiveness and reliability of our approach, demonstrating superior performance metrics and showcasing its exceptional potential for real-world fraud detection applications.

Phosphate-solubilizing fungus (PSF) - mediated phosphorous solubilization and validation through Artificial intelligence computation.

Phosphate-solubilizing fungus (PSF) strain alaromyces funiculosus was investigated for phosphorus solubilization, utilizing a range of pH levels and phosphate sources, followed by data confirmation through artificial intelligence modeling. T. funiculosus strain was exposed to five different phosphate sources [Ca3(PO4)2, FePO4, CaHPO4, AlPO4, and phytin] at different pH levels (4.5, 5.5, 6.5, 7.0, and 7.5). ANOVA, Pareto charts, and normal plots were used for analyzing the data. Artificial intelligence-based multilayer perceptron (MLP), random forest (RF) and extreme gradient boosting (XGBoost) models were used for data validation and prediction. Five-fold more phosphate (P) solubility by T. funiculosus was registered as compared to the control. The maximum soluble P was found at pH 4.5 (318324 ppb) and CaHPO4 (444045 ppb). Combination of phytin × 4.5 pH yielded the highest dissolved phosphorus (1537988 ppb), followed by 127458 ppb from the control × 4.5 pH. Pareto chart and normal plot analysis showedthe negative impact of pH (B), pH × F/C (fungus/control) × P-Source (ABC), and F/C (A) factor. Whereas pH × P-Source (AC) and P-Source (C) has positive impact on P solubility. The maximum R2 scores showed the order of RF (0.944) > MLP (0.938) > XGBoost (0.899). T. funiculosus strain has a grain potential for sustainable use for different types of phosphate sources. Application AI/ML models based on different performance metrics predicted the validated the attained results. In future research, it is recommended to check the efficacy of developed strategy under field conditions and to check the impact on soil and plant.

Development of New Electricity System Marginal Price Forecasting Models Using Statistical and Artificial Intelligence Methods

The System Marginal Price (SMP) is the cost of the last unit of electricity supplied to the grid, reflecting the supply–demand equilibrium and serving as a key indicator of market conditions. Accurate SMP forecasting is essential for ensuring market stability and economic efficiency. This study addresses the challenges of SMP prediction in Turkey by proposing a comprehensive forecasting framework that integrates machine learning, deep learning, and statistical models. Advanced feature selection techniques, such as Minimum Redundancy Maximum Relevance (mRMR) and Maximum Likelihood Feature Selector (MLFS), are employed to refine model inputs. The framework incorporates time series methods like Multilayer Perceptron (MLP), Long Short-Term Memory (LSTM), Bidirectional LSTM (Bi-LSTM), and Convolutional LSTM (ConvLSTM) to capture complex temporal patterns, alongside models such as Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), and Extreme Learning Machine (ELM) for modeling non-linear relationships. Model performance was evaluated using the Mean Absolute Percentage Error (MAPE) across regular weekdays, weekends, and public holidays. XGBoost combined with MLFS consistently achieved the lowest MAPE values, demonstrating exceptional accuracy and robustness. Among all of the models, XGBoost combined with MLFS consistently achieved the lowest MAPE values, demonstrating superior accuracy and robustness. The results highlight the inadequacy of traditional models like ARIMA and SARIMA in capturing non-linear and highly volatile patterns, reinforcing the necessity of using advanced techniques for effective SMP forecasting. Overall, this study presents a novel and comprehensive approach tailored for complex electricity markets, significantly enhancing predictive reliability by incorporating economic indicators and sophisticated feature selection methods.