Multilayer Perceptron Model Research Articles

Identification and evaluation of plasma exosome RNA biomarkers for non-invasive diagnosis of hepatocellular carcinoma using RNA-seq

BackgroundNon-invasive diagnostic methods, including medical imaging techniques and blood biomarkers such as alpha-fetoprotein (AFP), have been crucial in detecting hepatocellular carcinoma (HCC). However, imaging techniques are only effective for tumor size larger than 2 cm. AFP measurement remains unsatisfactory due to high rate of misdiagnosis and underdiagnosis. Therefore, new reliable biomarkers and better non-invasive diagnostic approach are necessary for HCC identification.MethodsThe differentially expressed genes were identified using multiple public RNA-seq data of liver tissues from healthy individuals and HCC patients including peritumoral and tumor tissues. The hub genes for HCC diagnosis were identified combining pathway enrichment analysis and protein–protein interaction network analysis. The performance of hub genes for non-invasive HCC diagnosis was analyzed in plasma of healthy individuals, HBV infected patients, and HCC patients based on exosomal RNA-seq data. A multi-layer perceptron (MLP) model based on exosomal hub genes was developed for non-invasive HCC diagnosis.ResultsThrough differential gene expression and pathway enrichment analysis on multiple public RNA-seq datasets, we first identified 30 dysregulated genes in HCC tissues. Protein-protein interaction analysis further narrowed down this list to 10 key genes: BRCA2, CDK1, MCM4, PLK1, DNA2, BLM, PCNA, POLD1, BRCA1 and FEN1. By further evaluation using additional public HCC tissue datasets, POLD1 and MCM4 were excluded from consideration as potential biomarkers due to their suboptimal performance. Notably, CDK1, FEN1, and PCNA gene were found to be significantly elevated in the plasma exosomes of HCC patients compared to non-HCC individuals, including those with HBV-infected hepatitis and healthy controls. The MLP model, based on three biomarkers, showed an area under the curve (AUC) of 0.85 and 0.84 in training and test dataset respectively, after adjusting for the covariates sex and age.ConclusionWe identified three key genes, CDK1, FEN1, and PCNA, as exosomal biomarkers for non-invasive diagnosis of HCC. The MLP model utilizing three biomarkers showed good differentiation between non-HCC individuals and HCC patients, which exhibits promising potential as a non-invasive diagnostic tool for detecting HCC. Additional validation with a larger sample size is essential to thoroughly assess the reliability of the biomarkers and the model’s performance.

Stochastic Learning Strategy with Varying Regularization Activation Layer-Based Diabetes Prediction Using Multilayer Perceptron

Diabetes makes people more susceptible to disorders like heart disease, kidney disease, Parkinson’s, cataracts, respiratory issues, and other illnesses. Numerous tests are employed to capture essential data for diagnosing diabetes, and the assessment is then used to decide the most appropriate course of action. Nevertheless, it is still challenging to identify the precise traits essential for the emergence of diabetes. Finding the critical factors that influence the target diabetes condition is the primary goal of this research. The Stochastic Learning Strategy with Varying Regularization Logistic Activation Multilayer Perceptron (MLP) model has been implemented with the Indian Diabetes patient dataset extracted from the UCI database repository. Data preprocessing was accomplished for the patients of the Indian diabetes dataset. This research tries to extract the correlation and distribution of all features from the Indian Diabetes Patient dataset. All classifiers demonstrate an accuracy of less than 80% when applied to the Indian Diabetes patient dataset with and without feature scaling, except the Random Forest classifier, which is found to have an accuracy of 78%, which is not the best. The Indian Diabetes patient dataset is fitted with MLP with various learning rates such as Constant learning, Constant momentum and Constant Nesterov momentum learning rate, Invasive Scaling, Invasive Scaling momentum and Invasive Scaling Nesterov momentum learning rate, Adaptive learning, Adaptive momentum and Adaptive Nesterov momentum learning rate for various activation layers like ReLU (Rectified Linear Unit), Identity, tanh and logistic activation layer and the performance is analyzed. The MLP with a logistic activation layer under the Adaptive momentum Learning rate strategy is examined with varying regularization parameter alpha to strengthen the prediction accuracy. The proposed Stochastic Learning Strategy with Varying Regularization Logistic Activation MLP model was further analyzed with the performance metrics precision, recall, FScore, runtime and accuracy and the experimental results exhibit 97% accuracy for MLP with logistic activation layer under Adaptive momentum Learning rate toward predicting the diabetes disease.

Derivation and Validation of Prediction of Preterm preeclampsia using machine learning algorithms.

To develop machine learning models for predicting preterm preeclampsia using the information available before 23 weeks' gestation. This was a secondary analysis of the Nulliparous Pregnancy Outcomes Study: Monitoring Mothers-to-Be cohort. We considered 131 features available before 23 weeks including maternal demographics, obstetrics and family history, social determinants of health, physical activity, nutrition, and early second-trimester ultrasound. Our primary outcome was preterm preeclampsia before 37 weeks. The dataset was randomly split into a training set (70%) and a validation set (30%). Machine learning models using glmnet, multilayer perceptron, random forest, XGBoost, and LightGBM models were developed. Using the machine learning approach that achieved the best area under the curves (AUC), we developed the final model. Further feature selection was conducted from the top 25 important features based on SHapley Additive exPlanations values. The performance of the final model was assessed using the validation dataset. Of 9,467 individuals, 219 (2.3%) had preterm preeclampsia. The AUC of the XGBoost model was the highest (AUC 0.749 [95% confidence interval 0.736-0.762]) compared with other models. Therefore, XGBoost was used to develop models using fewer variables. The XGBoost model with the 8 features (in order of importance: mean uterine artery pulsatility index in the early second trimester, chronic hypertension, pregestational diabetes, uterine artery notch, systolic and diastolic blood pressure in the first trimester, body mass index, and maternal age) was chosen as the final model as it had an AUC of 0.741 (95% confidence interval 0.730-0.752) which was not inferior to the original model (P=0.58). The final model in the validation dataset had an AUC of 0.779 (95% confidence interval 0.722-0.831). An online application of the final model was developed (https://kawakita.shinyapps.io/Preterm_preeclampsia/). Machine learning algorithms using information available before 23 weeks can accurately predict preterm preeclampsia before 37 weeks.

A Universal Method for Fingerprinting Multiplexed Bacteria: Evolving Pruned Sensor Arrays via Machine Learning-Driven Combinatorial Group-Specificity Strategy.

Array-based sensing technology holds immense potential for discerning the intricacies of biological systems. Nevertheless, developing a universal strategy for simultaneous identification of diverse types of multianalytes and meeting the diagnostic needs of a range of multiclassified clinical diseases poses substantial challenges. Herein, we introduce a combination method for constructing sensor arrays by assembling two types of group-specific elements. Such a method enables the rapid generation of a library of 100 sensing units, each with dual bacterial targeting capabilities. By employing a three-step screening strategy optimized by machine learning algorithms, various optimal five-element arrays were rapidly obtained for diverse clinical infectious models. Moreover, the pruned arrays successfully identified disparate mixing ratios and quantitative detection of clinically prevalent bacterial strains. Optimized through nine multiclassification algorithms, the top-performing multilayer perceptron (MLP) model demonstrated impressive recognition capabilities, achieving 100% accuracy for diagnosing clinical urinary tract infection (UTI) and 99.4% accuracy for clinical sepsis detection in the test models we collected. Such a combinatorial library construction and screening process should be standard and provides insights into successfully generating powerful high-recognition sensor elements and configuring them into highly discriminative mini-sensor arrays.

Fast prediction of key parameters in FEBA using the COSINE subchannel code and artificial neural network

Numerical techniques have emerged as an essential tool for operators and designers to preemptively acquire key parameters in accidents analysis. However, due to insufficient experience, it is difficult for them to obtain satisfactory numerical results. Moreover, the uncertainty analysis and quantification necessitate the simulation of a substantial number of samples, which requires a significant amount of computational time. Therefore, the development of a fast prediction model becomes imperative. In this work, a prediction model based on the in-house COSINE subchannel code and Multi-Head Perceptron (MHP) is developed. The COSINE subchannel code is employed to provide data sets for training neural networks. Firstly, the numerical results of COSINE subchannel code are compared with experimental data to ensure the accuracy of data sets. Secondly, input features for neural networks are selected by evaluating the impact of input parameters on numerical results, and a series of simulations is carried out to generate data sets. Then, a comparative analysis was conducted between the Multi-Layer Perceptron (MLP) and Support Vector Regression (SVR) models, and the MLP model performs better. Subsequently, the MLP was compared with the MHP, demonstrating the advantage of MHP model. Based on this, the predictions are conducted using the MHP model and the distribution of key parameters is compared with that obtained by COSINE subchannel code. The results illustrate that developed MHP model is an efficient tool for predicting key parameters during the reflooding phase.

Quasi-Newton Optimized Kolmogorov-Arnold Networks for Wind Farm Power Prediction

Having accurate and effective wind energy forecasting that can be easily incorporated into smart networks is important. Appropriate planning and energy generation predictions are necessary for these infrastructures. The production of wind energy is linked to instability and unpredictability. Wind energy forecasting has traditionally been performed using statistical models, but with the advent of artificial intelligence (AI), research indicates that AI is more accurate than the statical technique. In this study, the nominal power of six wind farms in China was predicted using Kolmogorov-Arnold Networks (KAN) and Multilayer Perceptron (MLP) models. KAN as an alternative to the conventional MLP, has the ability to handle problems with scalability, vanishing gradients, and interpretability associated with MLP. The KAN uses learnable B-Spline as activation functions prompting it to address the issues of the MLP. We employed the Radial Basis Function (RBF) with Gaussian kernels to approximate the 3-order B-spline basis. In most deep learning models stochastic gradient-based optimisation algorithms such as Adaptive Moment Estimation (ADAM) and Stochastic Gradient Descent (SGD) optimizer are mostly employed, a quasi-Newton optimization technique Limited-memory Broyden–Fletcher–Goldfarb–Shanno algorithm LBFGS was employed in this work to approximate the Hessian matrix and estimate the parameter space's curvature. Also, in the preprocessing of the data, the Interquartile Range (IQR) technique is used to handle outliers and a clustering-based K-Nearest Neighbor (KNN) imputer to handle missing values. Based on different sites, the KAN-LBFGS shows superior performance based on the performance evaluation metrics with site 5 achieving MSE of 0.0039, RMSE of 0.0622, MAE of 0.0352, and DC of 0.9468. The study highlights the importance of the model’s architecture, preprocessing and optimization techniques.

Deep Learning for Karolinska Sleepiness Scale Classification Based On Eye Aspect Ratio with SMOTE-Enhanced Data Balancing

This paper addresses the challenge of accurately classifying sleepiness levels based on the Karolinska Sleepiness Scale (KSS) using Eye Aspect Ratio (EAR) data, especially when class imbalance leads to biased predictions. The research proposes a deep learning framework that integrates a Multi-Layer Perceptron (MLP) with the Synthetic Minority Over-sampling Technique (SMOTE) to balance the training data. EAR features, representing eye closure patterns, are extracted from video frames, and SMOTE is applied to generate synthetic data for underrepresented sleepiness classes. By training the MLP model on this balanced dataset, the system achieves a 97.6% classification accuracy in distinguishing four distinct sleepiness levels based on the KSS, demonstrating its effectiveness in reducing prediction bias and managing class imbalance, both crucial for real-time drowsiness detection systems

Prediction of CO2 adsorption of biochar under KOH activation via machine learning

To effectively increase the CO2 adsorption capacity of biochar, the activation process is often an indispensable link. However, the introduction of an activation process poses challenges in clarifying the relationship between the characterization parameters of biochar, adsorption parameters, and CO2 adsorption capacity. Herein, a comprehensive dataset encompassing the CO2 adsorption data of KOH-activated biochar using a “two-sep method” was compiled. Subsequently, ridge regression, multi-layer perceptron, and random forest models were employed to predict its CO2 adsorption performance. To comprehensively explore the effects of activation conditions, physicochemical properties and adsorption parameters on CO2 adsorption capacities, partial dependence via Shapley additive explanation (SHAP) values analysis was conducted. The results demonstrate that the multilayer perceptron model exhibits the highest prediction accuracy with a test R2 value of 0.961. Additionally, it was found that the CO2 adsorption capacity of activated biochar is primarily determined by micropores and nitrogen-containing groups rather than total pore volume at low adsorption pressure (< 0.3 bar). Moreover, it increases significantly with decreasing average pore size, increasing pore volume, and increasing nitrogen content at low adsorption temperatures (< 20 °C). When the ratio of KOH to biochar is in the range of 1–2 and the activation temperature is ∼ 700 °C, activated biochar with high CO2 adsorption performance can be obtained. This study may provide valuable insights for the application of activated biochar in CO2 adsorption.

Liver Disease Prediction Using Deep Learning

Abstract: Liver, a crucial interior organ of the human body whose principal tasks are to eliminate generated waste produced by our organism, digest food, and preserve vitamins and energy materials. The liver disorder can cause various fatal diseases, including liver cancer. Early diagnosis, and treating the patients are compulsory to reduce the risk of those lethal diseases. we attempted to find a model which can predict the occurrence of liver disease in a given patient with the highest accuracy, based on different input factors. A dataset was chosen to train and test this model; Indian Liver Patient Dataset obtained . We implemented different deep learning algorithms (Multi- Layer Perceptron, Stochastic Gradient Descent, RestrictedBoltzmann) and filtered out the DL-based MLP (Multi-Layer Perceptron) model as the one providing the highest Accuracy, which was compared f or each model along with the Precision, Recall andf1 scores.

THE EFFECTS OF INTRINSIC ASPECTS OF WORK ON EMPLOYEE WELL-BEING

In recent years, increasing attention has been directed toward understanding the intrinsic aspects of work and their impact on employee well-being. This study aims to evaluate the influence of intrinsic work factors on employees' health, well-being, and job satisfaction. The research employs artificial neural network analysis, specifically a multilayer perceptron model, to establish relationships between intrinsic work factors (such as motivation, interest in the profession, and recognition) and employee outcomes, including health and well-being. The findings reveal that organizational motivation, interest in the profession, and recognition of effort are the most significant contributors to employee well-being, while other aspects, such as autonomy and task satisfaction, exert less influence. These results highlight the complexity of how intrinsic factors shape the overall work experience, suggesting a differentiated approach to human resource management. In conclusion, intrinsic work aspects positively impact employee well-being, but the influence varies across different factors, emphasizing the importance of fostering motivation, interest alignment, and recognition in the workplace.

Investigation and application of data balancing and combined discriminant model in rock burst severity prediction

In the development of intelligent rock burst prediction models, issues such as incomplete data coverage and data imbalance are frequently encountered. These issues may lead to risks of overfitting in predictive models, poor generalization capabilities, and increased bias, which in turn may result in misjudgments and unpredictable losses. To accurately predict rock burst disasters and mitigate or eliminate related threats, this paper proposes a composite prediction model that integrates Density-Based Nonlinear Resampling (DBNR)-Tomek Link data balancing algorithms with Bayesian Optimization (BO)-Multilayer Perceptron (MLP)-Random Forest (RF). Initially, this study collected and organized a total of 301 recorded rock burst disaster field observation data, covering various tectonic plates, engineering types, rock origins, rock textures, and rock burst types. Subsequently, from a data analysis perspective, we employed the PCA-SSA-K-means unsupervised clustering algorithm to delve into the underlying information contained within the data, thereby validating the rationality of categorizing rock bursts into four grades. Then, using the L2 norm to optimize the dimensionality of the indicators and supplementing with indicator importance ranking and hypothesis testing, we selected the maximum tangential stress of the surrounding rock, the ratio of the maximum tangential stress of the surrounding rock to the uniaxial compressive strength of the rock (stress coefficient), and the elastic energy index as the criteria for rock burst intensity grading. Following that, the DBNR-Tomek Link sampling method was applied to balance the sample data, optimizing the data sample ratio and ultimately expanding the sample size to 396, improving the proportion of data samples from 2:3:4:1 to 1:1:2:1, thereby enhancing the model’s generalization performance. Ultimately, a BO-MLP-RF composite prediction model was constructed based on Bayesian Optimization (BO), Multilayer Perceptron (MLP), and Random Forest (RF) algorithms, with the Bayesian Optimization method ensuring that the model fits the training data well and generalizes to the test data. The results of tenfold cross-validation demonstrated that the model’s accuracy is consistently around 92.5%, combining the training results of rock burst models with imbalanced datasets, proving that the MLP model, adept at modeling nonlinear data, and the RF model, skilled in modeling large-scale data, serve as basic classifiers. This demonstrates that the application of data balancing and combined discriminative model schemes have enhanced the model’s predictive performance and stability. The model is capable of providing high-accuracy, high-efficiency early warning monitoring services for rock burst phenomena in rock engineering, thereby ensuring engineering safety.

The deep continual learning framework for prediction of blast-induced overbreak in tunnel construction

Blast-induced overbreak, characterized by the excessive removal of rock mass beyond the planned tunnel profile, poses significant safety risks, increases costs, and causes project delays during tunneling. Traditional machine/deep learning models have been developed to predict overbreak. However, these models are often inadequate because they are static and lack the flexibility to adapt to new, real-world data continuously. This study addresses this limitation by introducing a novel data-driven approach based on deep continual learning. The primary objective is to develop an adaptable predictive model with the ability of continual learning, which is particularly advantageous in dynamic environments like tunnel blasting. To achieve this, a self-attention multi-layer perceptron (MLP) model for overbreak prediction, integrated with two continual learning strategies (elastic weight consolidation (EWC) and memory replay (MR)), is developed. This step enables the overbreak prediction model to possess the ability to continuously learn real-world scenarios and adapt to the dynamic environment of tunnel blasting. The findings show that the continuous MLP model, empowered by EWC and MR, demonstrates superior adaptability and accuracy in predicting overbreak. Compared with the standard MLP model, which achieves a predictive accuracy of 0.831, the continuous MLP model achieves a predictive accuracy of 0.845 on unseen data. The integration of EWC and MR strategies proves to be a pivotal factor in developing deep learning models for the dynamic task of predicting overbreak. The continual learning strategies ensure that the models remain adaptable and accurate over time, which is essential for practical applications in dynamic environments of tunnel blasting operations.

The Factors Influencing Safety Compliance Behavior Among New-Generation Construction Workers in China: A Safety Compliance Behavior–Artificial Neural Network Model Approach

Amid an aging workforce and labor shortages, this study investigates the key factors influencing construction workers’ safety compliance behavior (SCB). SCB is categorized into three distinct types: non-compliance behavior, general behavior, and compliance behavior. The study compares and analyzes the differences in influencing factors between the new generation and older generation of construction workers. By integrating the SCB framework with a multi-layer perceptron (MLP) model, this research develops a safety compliance behavior–artificial neural network (SCB-ANN) model. An enhanced method for optimizing connection weight (CW) is applied to identify the key determinants of SCB. The findings reveal that the SCB-ANN model offers superior predictive accuracy compared to a standard MLP model. Additionally, the refined CW method significantly improves the neural network’s interpretability. The analysis shows that organizational factors have a stronger influence on the new generation of construction workers (NGCWs), while individual factors play a more crucial role for the older generation (OGCWs). As a result, the study proposes tailored safety management measures for different worker groups to mitigate non-compliance behaviors, providing a robust foundation for future research and the development of safety management strategies.

Retrieval of cloud fraction using machine learning algorithms based on FY-4A AGRI observations

Abstract. Cloud fraction as a vital component of meteorological satellite products plays an essential role in environmental monitoring, disaster detection, climate analysis and other research areas. Random forest (RF) and multilayer perceptron (MLP) algorithms were used in this paper to retrieve the cloud fraction of AGRI (Advanced Geosynchronous Radiation Imager) on board the Fengyun-4A (FY-4A) satellite based on its full-disk level-1 radiance observation. Corrections have been made subsequently to the retrieved cloud fraction in areas where solar glint occurs using a correction curve fitted with sunglint angle as weight. The algorithm includes two steps: the cloud detection is conducted firstly for each AGRI field of view to identify whether it is clear sky, partly cloudy or overcast within the observation field. Then, the cloud fraction is retrieved for the scene identified as partly cloudy. The 2B-CLDCLASS-lidar cloud fraction product from the CloudSat and CALIPSO active remote sensing satellite is employed as the truth to assess the accuracy of the retrieval algorithm. Comparison with the operational AGRI level-2 cloud fraction product is also conducted at the same time. The results indicate that both the RF and MLP cloud detection models achieved high accuracy, surpassing that of operational products. However, both algorithms demonstrated weaker discrimination capabilities for partly cloudy conditions compared to clear-sky and overcast situations. Specifically, they tended to misclassify fields of view with low cloud fractions (e.g., cloud fraction = 0.16) as clear sky and those with higher cloud fractions (e.g., cloud fraction = 0.83) as overcast. Between the two models, RF exhibited higher overall accuracy. Both RF and MLP models performed well in cloud fraction retrieval, showing lower mean error (ME), mean absolute error (MAE) and root mean square error (RMSE) compared to operational products. The ME for both RF and MLP cloud fraction retrieval models was close to zero, while RF had slightly lower MAE and RMSE than MLP. During daytime, the high reflectance in sunglint areas led to larger retrieval errors for both RF and MLP algorithms. However, after correction, the retrieval accuracy in these regions improved significantly. At night, the absence of visible light observations from the AGRI instrument resulted in lower classification accuracy compared to daytime, leading to higher cloud fraction retrieval errors during nighttime.

Experimental analysis and low-damage machining strategy for composite ultrasonic vibration-assisted grinding of silicon carbide based on DA-MLP-NSGA-II algorithm

At present, because of the lack of ultrasonic composite vibration assisted grinding mechanism, neural network optimization algorithm (NNOA) is used to optimize the processing results. In NNOA, multi-layer perceptron (MLP) neural network model and non-dominated sorting genetic algorithm-II (NSGA-II) are very efficient and accurate methods. In this paper, based on the measurement and analysis of the specific ultrasonic vibration device, the CUVAG experiments on silicon carbide (SiC) ceramic were carried out to investigate the influence of processing parameters on the grinding forces, the ground surface roughness and morphology, and the subsurface damage. Then, the brittle-ductile removal behavior of hard-and-brittle materials could be revealed according to the above analysis. After that, MLP model and NSGA-II were utilized to predict and optimize the processing results in CUVAG. The results show that the grinding forces are basically constant, the surface quality deteriorates, and the subsurface damage increases with increased axial vibration amplitude and workpiece infeed speed, but all fluctuate with enlarged wheel speed, and turns at the inflection point of brittle-ductile transition with increased elliptic vibration amplitude. The fitting goodness R2 of the established MLP neural network prediction model is between 0.94 and 0.975, and the process parameters calculated by the NSGA-II optimization algorithm are verified. With optimized processing parameters, the grinding forces are reduced by about 13 %, the surface roughness is reduced to Ra0.037 μm (by 29 %), and the depth of subsurface damage is reduced by 68 %.

Machine learning-driven detection of anomalies in manufactured parts from resonance frequency signatures

ABSTRACT This study aims to enhance the detection and characterisation of anomalies in manufactured parts by integrating machine learning (ML) with resonance frequency spectra data. A key contribution of this work is the development of a novel Impulse Excitation Technique (IET)-based method that effectively evaluates material health and identifies subtle defects by leveraging numerous mathematical and physical metrics as input features. Three machine learning models – Random Forest (RF), K-Nearest Neighbor (KNN), and Multi-layer Perceptron (MLP) – were systematically compared to determine the most effective method for classifying defects, specifically focusing on healthy, cracked, and dimensionally deviated samples. Among these, the MLP model demonstrated the highest performance, achieving Receiver Operating Characteristic (ROC) values of 0.963, 0.901, and 0.942 for each class, respectively. Additionally, SHAP (SHapley Additive exPlanations) analysis showed anomalies were sensitive to specific resonance frequency metrics, improving prediction accuracy. Cracked samples exhibited slight peak broadening and negative peak shifts, while dimensionally deviated samples showed positive and negative shifts and missing peaks. Dimensional deviations were more pronounced than cracks, making them easier to identify and enhancing predictive accuracy.

Investigation of radiomic features on MRI images to identify extraprostatic extension in prostate cancer

Background and ObjectiveDetection of extraprostatic extension (EPE) preoperatively is of critical importance in the context of prostate cancer (PCa) management and outcomes. This study aimed to characterize the radiomic features of malignant prostate lesions based on multi-paramagnetic magnetic resonance imaging (mpMRI). MethodsWe analyzed 20 patients who underwent mpMRI followed by radical prostatectomy. Two experienced radiologists manually segmented the 3D lesions using the T2-weighted (T2WI) and Apparent Diffusion Coefficient (ADC) imaging sequences. A total of 210 radiomic features were extracted from each lesion. We used the Recursive Feature Elimination with Cross-Validation to select key features. Using the selected radiomic features, we developed a Multilayer Perceptron (MLP) neural network to classify the EPE and non-EPE lesions. The pathology results were accepted as gold standard for EPE. We measured the performance of the classifier, calculating the area-under-curve (AUC), sensitivity, and specificity. ResultsA total of 25 lesions were segmented, including 12 lesions with EPE and 13 lesions without EPE, based on the pathology reports. We selected 18 radiomic features (18/210). The MLP classifier using these features provided a good sensitivity (0.75), specificity (0.79), and AUC of 0.82, 95 % CL [0.59 - 0.96] in identifying the EPE lesions. ConclusionsThis pilot study presents 18 radiomic features derived from T2-weighted and ADC images and demonstrates their potential in the preoperative prediction of EPE in PCa using an MLP model.

Establishment of an MRI-based radiomics model for distinguishing between intramedullary spinal cord tumor and tumefactive demyelinating lesion

ObjectiveDifferentiating intramedullary spinal cord tumor (IMSCT) from spinal cord tumefactive demyelinating lesion (scTDL) remains challenging with standard diagnostic approaches. This study aims to develop and evaluate the effectiveness of a magnetic resonance imaging (MRI)-based radiomics model for distinguishing scTDL from IMSCT before treatment initiation.MethodsA total of 75 patients were analyzed in this retrospective study, comprising 55 with IMSCT and 20 with scTDL. Radiomics features were extracted from T1- and T2-weighted imaging (T1&T2WI) scans upon admission. Ten classification algorithms were employed: logistic regression (LR); naive bayes (NaiveBayes); support vector machine (SVM); k nearest neighbors (KNN); random forest (RF); extra trees (ExtraTrees); eXtreme gradient boosting (XGBoost); light gradient boosting machine (LightGBM); gradient boosting (GradientBoosting); and multi-Layer perceptron (MLP). The performance of the optimal model was then compared to radiologists' assessments.ResultsThis study developed 30 predictive models using ten classifiers across two imaging sequences. The MLP model with two sequences (T1&T2WI) emerged as the most effective one, showing superior accuracy in MRI analysis with an area under the curve (AUC) of 0.991 in training and 0.962 in testing. Moreover, statistical analyses highlighted the radiomics model significantly outperformed radiologists' assessments (p < 0.05) in distinguishing between IMSCT and scTDL.ConclusionWe present an MRI-based radiomics model with high diagnostic accuracy in differentiating IMSCT from scTDL. The model’s performance was comparable to junior radiologists, highlighting its potential as an effective diagnostic aid in clinical practice.

A Cutting‐Edge Hybrid Deep Learning Technique with Low Rank Approximation for Attacks Classification on IoT Traffic Data

ABSTRACTNetwork security is experiencing huge challenges as network attacks on traffic data become more frequent and sophisticated. In this paper, we employ hybrid deep learning models and low‐rank approximation to present a novel method for multi‐label categorization of network assaults on traffic data. Our suggested solution, LR‐CNN‐MLP, consists of three models a multi‐layer perceptron (MLP), a hybrid convolutional neural network (CNN), and a low‐rank approximation model. While the CNN and MLP models extract features and categorize data, respectively, the low‐rank approximation model reduces the input's dimensionality. Overall, by combining hybrid models and low‐rank approximation, our proposed LR‐CNN‐MLP approach provides a promising solution for multi‐label categorization of network attacks on traffic data. LR‐CNN‐MLP achieves the highest results of performance metrics such as 0.944 precision, 0.979 recall, 0.961 F1‐score and 98.17 accuracy.

Prediction of Intraday Electricity Supply Curves

The electricity market in Spain, as in many European countries, is organized into daily, intraday, and reserve markets. This project aims to predict the supply curves in the Spanish intraday market that have six sessions with different horizons of application, using information from the market itself. To achieve this, we approximate these curves using a non-uniform grid of points and evaluate the quality of these approximations with a weighted distance, both based on empirical market data. We employ neural network models, including multilayer perceptrons (MLPs), convolutional neural networks (CNNs), long short-term memory (LSTM), bidirectional LSTM (BiLSTM), and a Transformer network alongside a naive model for benchmarking. The MLP and CNN models demonstrated significant improvements in predicting these supply curves for the six market sessions.