Multilayer Perceptron Research Articles

Application of a Physics-Informed Convolutional Neural Network for Monitoring the Temperature Fields in High-Temperature Gas Reactors

This work presents current advances in applying a physics-informed convolutional neural network (CNN) to evaluate temperature distributions in advanced reactors. Our goal is to demonstrate that the CNN can reconstruct temperature fields within the solid region of a prismatic fuel assembly in a high-temperature gas reactor (HTGR) with sensor data available in only a few cooling channels. Before that, we showcase the superior performance of the physics-informed CNN in comparison to a purely data-driven multilayer perceptron (MLP), considering a canonical heated channel setup. This analysis shows the advantages of our approach and justifies its choice. The datasets employed here are obtained upon numerical simulations performed with codes under the Nuclear Energy Advanced Modeling and Simulation program. This work is important, as industry experience indicates that the assembly material in HTGR concepts is prone to large thermal-mechanical loads nearing operational limits. This makes it crucial to characterize peak temperatures and their distributions near hot spots. Modern thermocouples are unreliable in these types of harsh environments because of the high neutron fluxes and elevated temperatures involved. The CNN-based field reconstruction represents an attractive solution, enabling sensor arrays in less aggressive locations and augmenting indirect predictions for less accessible regions. The results show that the CNN reduces prediction errors by orders of magnitude in comparison to the MLP, considering the simple yet well-representative heated channel case. In the case of the HTGR fuel assembly, the CNN can successfully reconstruct temperature fields over various cooling regimes. Furthermore, we also explore the algorithm’s ability to detect abnormalities. Interestingly, the CNN proves it has the capacity to detect blockage in one of the noninstrumented cooling channels.

Smart energy management: Process structure-based hybrid neural networks for optimal scheduling and economic predictive control in integrated systems

Integrated energy systems (IESs) are complex prosumers consisting of diverse operating units spanning multiple domains. The tight integration of these units results in varied dynamic characteristics and intricate nonlinear process interactions, making detailed dynamic modeling and successful operational optimization challenging. To address these concerns, we propose a process structure-based hybrid time-series neural network (NN) surrogate to predict the dynamic performance of IESs across multiple time scales. This neural network-based modeling approach develops time-series multi-layer perceptrons (MLPs) for the operating units and integrates them with prior process knowledge about system structure and fundamental dynamics. This integration forms three hybrid NNs—long-term, slow, and fast MLPs—that predict the entire system dynamics across multiple time scales. Leveraging these MLPs, we design an NN-based scheduler and an NN-based economic model predictive control (NEMPC) framework to meet global operational requirements: rapid electrical power responsiveness to operators’ requests, adequate cooling supply to customers, and increased system profitability, while addressing the dynamic time-scale multiplicity present in IESs. The proposed day-ahead scheduler is formulated using the ReLU network-based MLP, which effectively represents IES performance under a broad range of conditions from a long-term perspective. The scheduler is then exactly recast into a mixed-integer linear programming problem for efficient evaluation. The real-time NEMPC, based on slow and fast MLPs, comprises two sequential distributed control agents: a slow NEMPC for the cooling-dominant subsystem with slower transient responses and a fast NEMPC for the power-dominant subsystem with faster responses. These agents collaborate in the decision-making process to achieve dynamic synergy in real time while reducing computational costs. Extensive simulations demonstrate that the developed scheduler and NEMPC schemes outperform their respective benchmark scheduler and controller by about 25% and 40%. Together, they enhance overall system performance by over 70% compared to benchmark approaches.

Machine learning-based sales forecasting during crises: Evidence from a Turkish women's clothing retailer.

Retail involves directly delivering goods and services to end consumers. Natural disasters and epidemics/pandemics have significant potential to disrupt supply chains, leading to shortages, forecasting errors, price increases, and substantial financial strains on retailers. The COVID-19 pandemic highlighted the need for retail sectors to prepare for crisis impacts on sales forecasts by regularly assessing and adjusting sales volumes, consumer behavior, and forecasting models to adapt to changing conditions. This study explores strategies for adapting sales forecasts and retail approaches in response to such crises. By employing different machine learning (ML) methods, we analyze consumer behavior changes and sales impacts across various product categories, including bottom wear, top wear, one piece, accessories, outwear, and shoes during the COVID-19 pandemic. The gradient boosting and CatBoost algorithms excelled in product groups with significant sales changes during the pandemic. The Multi-Layer Perceptron (MLP) algorithm performed well in low-volume categories like accessories and footwear. Meanwhile, MLP, LightGBM, and XGBoost were effective in medium-volume categories such as outerwear and underwear. The findings highlight the efficacy of these models in adapting sales forecasts to crisis conditions, offering a practical approach to enhancing retail resilience against future disruptions. This study offers an effective approach for adapting sales forecasting to shifting consumer behaviors during crises.

Small-data-trained model for predicting nitrate accumulation in one-stage partial nitritation-anammox processes controlled by oxygen supply rate.

Nitrate (NO3--N) accumulation is the biggest obstacle for wastewater treatment via partial nitritation-anammox process. Dissolved oxygen (DO) control is the most used strategy to prevent NO3--N accumulation, but the performance is usually unstable. This study proposes a novel strategy for controlling NO3--N accumulation based on oxygen supply rate (OSR). In comparison, limiting the OSR is more effective than limiting DO in controlling NO3--N accumulation through mathematical simulation. A laboratory-scale one-stage partial nitritation-anammox system was continuously operated for 135 days, which was divided into five stages with different OSRs. A novel deep learning model integrating Gated Recurrent Unit and Multilayer Perceptron was developed to predict NO3--N accumulation load. To tackle with the general obstacle of limited environmental samples, a generic evaluation was proposed to optimise the model structure by leveraging predictive performance and overfitting risk. The developed model successfully predicted the NO3--N accumulation in the system ten days in advance, showcasing its potential contribution to system design and performance enhancement.

Artificial neural networks classification of s-band absorption performance in eco-friendly microwave absorbers

Microwave absorbers are essential for applications such as radar stealth and electromagnetic compatibility. Nevertheless, traditional materials encounter obstacles related to cost and sustainability, which has led to the exploration of new options such as materials derived from agricultural waste. This study focuses on the classification challenge of evaluating the absorption performance of eco-friendly microwave absorbers in the S-band (2 to 4 GHz) frequency. Three multilayer perceptron (MLP) algorithms, namely levenberg marquardt (LM), resilient backpropagation (RBP) and scale conjugate gradient (SCG) are assessed for classification accuracy. The dataset consists of 135 absorption performance values of microwave absorbers that were taken from experimental measurements using the naval research laboratory (NRL) arch free. The MLP algorithms will be divided into three divisions, which are training, validation and testing, evaluating important criteria such as accuracy, precision, sensitivity and specificity. The performance of three types of algorithms will be compared using two basic inputs: the absorption values and the single slot sizes. The RBP algorithm achieved 100% accuracy, and a lower mean squared error (MSE) of 0.02500 compared to the LM and SCG. This study provides valuable insights for designing better microwave absorbers and highlights the commercial potential of agricultural waste materials in such applications.

Multi-channel spatio-temporal graph attention contrastive network for brain disease diagnosis.

Dynamic brain networks (DBNs) can capture the intricate connections and temporal evolution among brain regions, becoming increasingly crucial in the diagnosis of neurological disorders. However, most existing researches tend to focus on isolated brain network sequence segmented by sliding windows, and they are difficult to effectively uncover the higher-order spatio-temporal topological pattern in DBNs. Meantime, it remains a challenge to utilize the structure connectivity prior in the DBNs analysis. To address these problems, we propose a multi-channel spatio-temporal graph attention contrastive network for DBNs analysis. Specifically, we first construct dynamic brain functional networks from fMRI data with sliding windows, and embed the structural connectivity derived from diffusion tensor imaging (DTI) to the dynamic functional connectivity graph representation to construct multi-modal brain network. Second, we develop a multi-channel spatial attention contrastive network to extract topological features from the brain network within each time window. This network incorporates an intra-window graph contrastive constraint to enhance the discriminative ability of the extracted features. Moreover, temporal dependencies across windows are captured by integrating feature embeddings through a self-attention mechanism, and the inter-window recurrent contrastive constraint is devised to extract higher-order spatio-temporal topological features. Finally, a multi-layer perceptron (MLP) is used to classify the brain networks. Experiments on epilepsy and ADNI datasets show that our method outperforms several state-of-the-art approaches in diagnosing performance, and it provides discriminative graph features for related brain diseases.

CLPr_in_ML: Cleft Lip and Palate Reconstructed Features with Machine Learning

Background: Cleft lip and palate are two of the most common craniofacial congenital malformations in humans. It influences tens of millions of patients worldwide. The hazards of this disease are multifaceted, extending beyond the obvious facial malformation to encompass physiological functions, oral health, psychological well-being, and social aspects. Objective: The primary objective of our study is to demonstrate the importance of imaging in detecting cleft lip and palate. By observing the morphological and structural abnormalities involving the lip and palate through imaging methods, this study aims to establish imaging as the primary diagnostic approach for this disease. Methods: In this work, we proposed a novel model to analyze unilateral complete cleft lip and palate after velopharyngeal closure and non-left lip and palate patients from the Department of Stomatology of Xuzhou First People's Hospital, Conical Beam CT (CBCT) images in silicon. In order to demonstrate the generalization, the simulated dataset was constructed using the random disturbance factor, which is from the actual dataset. We extracted several raw features from CBCT images in detail. Then, we proposed a novel feature reconstruction method, including six types of reconstructed factors, to reconstruct the existing features. Then, the reconstructed features weretrained with machine learning algorithms. Finally, the testing and independent data model was utilized to analyze the performance of this work. Results: By comparing different operator features, the min operator, max operator, average operator, and all operators can achieve good performances in both the testing set and the independent set. Conclusion: With the different operator features, the majority of classification models, including Gradient Boosting, Hist Gradient Boosting, Multilayer Perceptron, lightGBM, and broadened learning, classification algorithms can get the well-performances in the selected reconstructed feature operators.

Detection of sleep apnea using only inertial measurement unit signals from apple watch: a pilot-study with machine learning approach

PurposeDespite increased awareness of sleep hygiene, over 80% of sleep apnea cases remain undiagnosed, underscoring the need for accessible screening methods. This study presents a method for detecting sleep apnea using data from the Apple Watch’s inertial measurement unit (IMU).MethodsAn algorithm was developed to extract seismocardiographic and respiratory signals from IMU data, analyzing features such as breathing and heart rate variability, respiratory dips, and body movements. In a cohort of 61 adults undergoing polysomnography, we analyzed 52,337 30-second epochs, with 12,373 (23.6%) identified as apnea/hypopnea episodes. Machine learning models using five classifiers (Logistic Regression, Random Forest, Gradient Boosting, k-Nearest Neighbors, and Multi-layer Perceptron) were trained on data from 41 subjects and validated on 20 subjects.ResultsThe Random Forest classifier performed best in per-epoch respiratory event detection, achieving an AUC of 0.827 and an F1 score of 0.572 in the training group, and an AUC of 0.831 and an F1 score of 0.602 in the test group. The model’s per-subject predictions strongly correlated with the apnea-hypopnea index (AHI) from polysomnography (r = 0.93) and identified subjects with AHI ≥ 15 with 100% sensitivity and 90% specificity.ConclusionUtilizing the widespread availability of the Apple Watch and the low power requirements of the IMU, this approach has the potential to significantly improve sleep apnea screening accessibility.

Harnessing temporal patterns in administrative patient data to predict risk of emergency hospital admission.

Unplanned hospital admissions are associated with worse patient outcomes and cause strain on health systems worldwide. Primary care electronic health records (EHRs) have successfully been used to create prediction models for emergency hospitalisation, but these approaches require a broad range of diagnostic, physiological, and laboratory values. In this study, we aimed to capture temporal patterns of patient activity from EHR data and evaluate their effectiveness in predicting emergency hospital admissions compared with conventional methods. In this retrospective observational study, we used the Secure Anonymised Information Linkage databank to extract temporal patterns of primary care activity from undifferentiated electronic health record timestamp data for 1·37 million patients in Wales aged 18-80 years with at least one recorded Read code between the years 2016 and 2018. Using Gaussian mixture modelling we grouped patients into distinct temporal clusters, performed a three-stage validation of our approach and calculated the risk of emergency hospital admission for each temporal cluster group. Finally, these temporal clusters were combined with five administrative variables and incorporated into four emergency hospital admission prediction models (logistic regression, naive Bayes, XGBoost, and multilayer perceptron [MLP]) and compared with a more traditional, but data-intensive, modelling technique. The primary outcome was emergency hospital admission as the next health-care event. Six distinct temporal cluster patterns of primary care EHR activity were identified, associated with varying risks of future emergency hospital admission risk. These patterns were visually interpretable, repeatable at a population-level, and clinically plausible. The best emergency hospital admission prediction model (MLP) achieved an area under the receiver operating characteristic (AUROC) of 0·82 and precision of 0·94 in regional cohorts. In external validation in regional cohorts, similar model performance was observed (AUROC 0·82 and precision 0·92). This model also matched the performance of a more complex model (extended feature model) requiring 33 clinical parameters (AUROC 0·82 vs 0·83; precision 0·94 vs 0·90) for the same task on the same dataset. We developed a novel machine learning pipeline that extracts interpretable temporal patterns from simple representations of EHR data and can be incorporated into emergency hospital admission predictors. This framework might enable more rapid development of parsimonious clinical prediction models. UKRI CDT in AI for Healthcare, UKRI Turing AI Fellowship, NIHR Imperial Biomedical Research Centre, and Research Capability Funding.

Denigration analysis of Twitter data using cyclic learning rate based long short-term memory

Technological innovation has given rise to a new form of bullying, often leading to significant harm to one's reputation within social circles. When a single person becomes target to animosity and harassment in a cyberbullying incident, it is termed as denigration. Many different cyberbullying detection techniques are carried out to counter this, concentrating on word-based data and user account features only. The main objective of this research is to enhance the learning rate of long short-term memory (LSTM) using cyclic learning rate (CLR). Therefore, in this research, cyberbullying in social media is detected by developing a framework based on LSTM-CLR which is more stable for enhancing classification accuracy without the need for multiple trials and modifications. The effectiveness of the suggested LSTM-CLR is assessed for identifying cyberbullying using Twitter data. The attained results show that the proposed LSTM-CLR obtains 82% accuracy, 80% precision, 83% recall and 81% F-measure in the classification of cyberbullying tweets, which is superior when compared with the existing multilayer perceptron (MLP) and bidirectional encoder representations from transformers (BERT) models.

Transformer with token attention and attribute prediction for image captioning

Recently, Vision Transformers (ViTs) have become the mainstream models in image captioning tasks. ViTs take all image tokens as inputs to extract visual features, which may cause concerns about worthless tokens, and meanwhile lead to a huge amount of computation. This paper proposes a novel token reduction module to remedy this drawback. Specifically, the module employs ViTs to embed the input tokens, and adaptively learns informative visual tokens in way of token attention on the channel-spatial granularity. Furthermore, an attribute prediction module is designed to strengthen the relationship between vision and language. Technically, the attribute prediction is achieved via a classifier in form of Multi-Layer Perceptron (MLP). Both the visual representations and attribute representations are obtained by Transformers, which are then combined as the input of the Transformer decoder for caption generation. All of the modules are constructed in an encoder–decoder framework and support the end-to-end learning. Experiment results have shown that our approach can effectively reduce the computational cost of ViTs while maintaining comparable performance on the MS COCO and NoCaps datasets. For example, by pruning more than 70% of the input tokens, our approach greatly reduces GFLOPs by 41% ∼ 47%, while preserving its accuracy of a 142.1 CIDEr score on the MS COCO dataset.

An Approach to Forecast Quality of Water Effectively Using Machine Learning Algorithms

Background:: The quality of water directly or indirectly impacts the health and environmental well-being. Data about water quality can be evaluated using a Water Quality Index (WQI). Computing WQI is a quick and affordable technique to accurately summarise the quality of water. Objective:: The objective of this study is to find strategies for data preparation to categorize a dataset on the water quality in two remote Indian villages in different geographic locations, to predict the quality of water, and to identify low-quality water before it is made accessible for human consumption. Methods:: To accomplish this task, four water quality features Nitrate, pH, Residual Chlorine, and Total Dissolved Solids which are crucial for human consumption, are considered to dictate the quality of water. Methods used in handling these features include five steps that are data preprocessing with min-max normalization, finding WQI, using feature correlation to identify parameter importance with WQI, application of supervised machine learning regression models such as Random Forest (RF), Multiple Linear Regression (MLR), Gradient Boosting (GB) and Support Vector Machine (SVM) for WQI prediction. Then, a variety of machine learning classification models, including K-Nearest Neighbour (KNN), Support Vector Classifier (SVC), and Multi-layer Perceptron (MLP), are ensembled with Logistic Regression (LR), acting as a meta learner, to create a stack ensemble model classifier to predict the Water Quality Class (WQC) more accurately. Results:: The examination of the testing model revealed that RF regression and MLR algorithms performed best in predicting the WQI with mean absolute error (MAE) of 0.003 and 0.001 respectively. Mean square error (MSE), root mean square error (RMSE), R squared (R2), and Explained Variance Score (EVS) findings are 0.002,0.005,0.988 and 0.998 respectively with RF while 0.001,0.031,0.999 and 0.999 respectively with MLR. Meanwhile, for predicting WQC, the stack model classifier showed the best performance with an Accuracy of 0.936, F1 score of 0.93, and Matthews Correlation Coefficient (MCC) of 0.893 for the dataset of Lalpura and Accuracy of 0.991, F1 Score of 0.991 and MCC of 0.981 respectively for the dataset of Heingang. Conclusion:: This study explores a method for predicting water quality that combines easy and feasible water quality measurements with machine learning. The stack model classifier performed best for multiclass classification, according to this study. To ensure that the highest quality of water is given throughout the year, information from this study will motivate researchers to look into the underlying root causes of the quality variations.

Green solvent-based extraction of cellulose from hemp bast fibers: From treatment efficacy to characterizations and optimization.

This study assessed the efficacy of deep eutectic solvents (DESs) as green solvents for cellulose extraction from hemp bast fiber. Three DES mixtures were applied, and the combination of choline chloride and glycerol was selected for further experimentation due to its superior performance. The impact of four key parameters- treatment time, treatment temperature, DES-to-hemp ratio, and glycerol-to-choline chloride ratio was analyzed using Central-Composite Design (CCD) within response surface methodology (RSM) to optimize cellulose extraction. The optimal conditions achieved cellulose: lignin ratio of 70.817. Machine learning (ML) algorithms, including Multilayer Perceptron, Instance-Based Learner (IBK), Random Committee, Random Forest and Random Tree were conducted to predict the extraction process based on the RSM results. The results showed that the Random Tree demonstrated superiority by providing a predicted R2 value of 0.8548. Various characterization techniques such as SEM, FTIR, TGA, and XRD confirmed the removal of impurities. TGA and XRD results indicate a crystallinity index of 81.4% and an increase in cellulose yield from 58.108% in untreated hemp to 69.731% in the h-12 sample, respectively, consistent with RSM findings. These findings indicate the high potential of DES as a green, cost-effective, and environmentally friendly solvent for cellulose extraction from hemp.

Free-Phish: detecting phishing websites hosted on free web hosting domains

Phishing attacks have changed over time, and a family of phishing websites hosted over free web hosting domains (FHDs) are increasing. These can be created and maintained at scale with minimal effort while also effectively evading prevalent anti-phishing detection and resisting website takedowns. Their coverage time is also slower; hence, it's challenging for anti- phishing tools to catch them. From August 2022 to April 2023, more than 5,000 of these FHD URLs were observed on phish- tank.com and openphish.com by 41 different FHD providers. This paper investigates a family of phishing websites hosted on free web hosting domains (FHDs). We have developed an updated dataset containing both FHD and non-FHD URLs. While we leveraged some pre-existing features, we also introduced two novel features unique to this study. This work provides the first comprehensive study focused on distinguishing FHD and non-FHD URLs, filling a notable gap in the literature. Our dataset comprises 4,922 recently collected FHD URLs and 5,010 non-FHD URLs. By employing a multilayer perceptron neural network with 32 features, we achieved an impressive accuracy rate of 96.5% for identifying FHD URLs. Additionally, using a random forest classifier, we achieved an accuracy of 97.7% for non-FHD URLs.

Research on supercritical CO2-based filled type U-pipe solar evacuated tubular collector: A thermodynamically parametric optimization and implementation machine learning modeling

Using supercritical CO2 as a working fluid in a heat and power system operating on solar energy is an advantageous approach that achieves higher efficiency, carbon capture, and emissions reduction. Therefore, the present work focused on researching the effectiveness of developing a supercritical CO2-based U-pipe solar evacuated tubular collector (ETC) by a filled layer that includes contributions to thermodynamic parametric optimization and machine learning modeling to forecast the outlet temperature intelligently. Results indicated that the filled type increase the outlet temperature by 24.29 % compared with the conventional type. The evaluation of the optimization research findings shows that an optimal mass flow rate of 8.0 kg/hr achieves a maximum exergy efficiency of 13.70 % to 14.51 %. Furthermore, the results recommended that arranging 13 tubes with a length of 1800 mm is optimal for exergy efficiency within a mass flow rate range of 6.0–12.0 kg/hr. Based on machine learning implementations, Gradient Boosting, Support Vector Machine, and Multi-Layer Perceptron neural networks can intelligently predict with R2|Tout = 1.000, 0.997, and 0.995, respectively. A polynomial-based GMDH/ML with R2|Tout = 0.986 was presented to intelligently forecast the outlet temperature using five input (Tamb,ṁ,ηopt,Ltube,It) parameters.

Airport infrastructure and runway precision aids for forecasting flight arrival delays

Recent research has concentrated on using machine learning approaches to forecast flight delays. The majority of prior prediction algorithms were based on simple and standard attributes collected from the database from which the data were pulled. This article is the first attempt to propose novel features linked to airport capacity and infrastructure. The total runways, the total runway intersections, the longest runway length, the shortest runway length, the runway precision rate, the total terminals, and the total gates were all examined. In this paper, we suggest an optimized multilayer perceptron to predict flight arrival retards implementing data for domestic flights operated in United States airports. We employed data normalization, sampling techniques, and hyper-parameter tuning to strengthen the reliability of the suggested model. The experimental findings demonstrated that data normalization, sampling approaches, and Bayesian optimization produced the most accurate model with 92.49% accuracy. The achievements of the study were compared to other benchmark research from literature. The time complexity for the proposed model was computed and presented at the end of the investigation.

An integrated feature extraction framework of linear multi-layer perceptron to reduce computation complexity for remaining useful life prediction

An integrated feature extraction framework of linear multi-layer perceptron to reduce computation complexity for remaining useful life prediction

Estimating travel time in transport network with a combined multi-attributed graph convolutional neural network and multilayer perceptron model

Estimating travel time in transport network with a combined multi-attributed graph convolutional neural network and multilayer perceptron model

HEDDI-Net: heterogeneous network embedding for drug-disease association prediction and drug repurposing, with application to Alzheimer’s disease

BackgroundThe traditional process of developing new drugs is time-consuming and often unsuccessful, making drug repurposing an appealing alternative due to its speed and safety. Graph neural networks (GCNs) have emerged as a leading approach for predicting drug-disease associations by integrating drug and disease-related networks with advanced deep learning algorithms. However, GCNs generally infer association probabilities only for existing drugs and diseases, requiring network re-establishment and retraining for novel entities. Additionally, these methods often struggle with sparse networks and fail to elucidate the biological mechanisms underlying newly predicted drugs.MethodsTo address the limitations of traditional methods, we developed HEDDI-Net, a heterogeneous embedding architecture designed to accurately detect drug-disease associations while preserving the interpretability of biological mechanisms. HEDDI-Net integrates graph and shallow learning techniques to extract representative diseases and proteins, respectively. These representative diseases and proteins are used to embed the input features, which are then utilized in a multilayer perceptron for predicting drug-disease associations.ResultsIn experiments, HEDDI-Net achieves areas under the receiver operating characteristic curve of over 0.98, outperforming state-of-the-art methods. Rigorous recovery analyses reveal a median recovery rate of 73% for the top 100 diseases, demonstrating its efficacy in identifying novel target diseases for existing drugs, known as drug repurposing. A case study on Alzheimer's disease highlighted the model's practical applicability and interpretability, identifying potential drug candidates like Baclofen, Fluoxetine, Pentoxifylline and Phenytoin. Notably, over 40% of the predicted candidates in the clusters of commonly prescribed clinical drugs Donepezil and Galantamine had been tested in clinical trials, validating the model's predictive accuracy and practical relevance.ConclusionsHEDDI-NET represents a significant advancement by allowing direct application to new diseases and drugs without the need for retraining, a limitation of most GCN-based methods. Furthermore, HEDDI-Net provides detailed affinity patterns with representative proteins for predicted candidate drugs, facilitating an understanding of their physiological effects. This capability also supports the design and testing of alternative drugs that are similar to existing medications, enhancing the reliability and interpretability of potential repurposed drugs. The case study on Alzheimer's disease further underscores HEDDI-Net's ability to predict promising drugs and its applicability in drug repurposing.

Differential Gene Expression Analysis and Machine Learning identified Structural, TFs, Cytokine and Glycoproteins, including SOX2, TOP2A, SPP1, COL1A1, and TIMP1 as potential drivers of Lung Cancer

Background Lung cancer is a primary global health concern, responsible for a considerable portion of cancer-related fatalities worldwide. Understanding its molecular complexities is crucial for identifying potential targets for treatment. The goal is to slow disease progression and intervene early to prevent the development of advanced lung cancer cases. Hence, there's an urgent need for new biomarkers that can detect lung cancer in its early stages. Methods: The study conducted RNA-Seq analysis of lung cancer samples from the publicly available SRA database (NCBI SRP009408), including both control and tumour samples. The genes with differential expression between tumour and healthy tissues were identified using R and Bioconductor. Machine learning (ML) techniques, Random Forest, Lasso, XGBoost, Gradient Boosting, and Elastic Net were employed to pinpoint significant genes followed by classifiers, Multilayer Perceptron (MLP), Support Vector Machines (SVM), and k-Nearest Neighbors (k-NN). Gene ontology and pathway analyses were performed on the significant differentially expressed genes (DEGs). The top genes from DEG and machine learning analyses were combined for protein-protein interaction (PPI) analysis, identifying 10 hub genes essential for lung cancer progression. Results: The integrated analysis of ML and DEGs revealed the significance of specific genes in lung cancer samples, identified the top five upregulated genes (COL11A1, TOP2A, SULF1, DIO2, MIR196A2) and the top five downregulated genes (PDK4, FOSB, FLYWCH1, CYB5D2, MIR328), along with their associated genes implicated in pathways or co-expression networks were identified. Among the various algorithms employed, Random Forest and XGBoost proved effective in identifying common genes, underscoring their potential significance in lung cancer pathogenesis. The MLP exhibited the highest accuracy in classifying samples using all genes. Additionally, the protein-protein interaction (PPI) analysis identified 10 hub genes that are pivotal in lung cancer pathogenesis: COL1A1, SOX2, SPP1, THBS2, POSTN, COL5A1, COL11A1, TIMP1, TOP2A, and PKP1. Conclusion The study contributes to the early prediction of lung cancer by identifying potential biomarkers that could enhance early diagnosis and pave the way for practical clinical applications in the future. Integrating DEGs and machine learning-derived significant genes for PPI analysis offers a robust approach to uncovering critical molecular targets for lung cancer treatment.