Multi-class Classification Problem Research Articles

A preordonance-based decision tree method and its parallel implementation in the framework of Map-Reduce

In supervised classification, decision trees are one of the most popular learning algorithms that are employed in many practical applications because of their simplicity, adaptability, and other perks. The development of effective and efficient decision trees remains a major focus in machine learning. Therefore, the scientific literature provides various node splitting measurements that can be utilized to produce different decision trees, including Information Gain, Gain Ratio, Average Gain, and Gini Index. This research paper presents a new node splitting metric that is based on preordonance theory. The primary benefit of the new split criterion is its ability to deal with categorical or numerical attributes directly without discretization. Consequently, the Preordonance-based decision tree” (P-Tree) approach, a powerful technique that generates decision trees using the suggested node splitting measure, is developed. Both multiclass classification problems and imbalanced data sets can be handled by the P-Tree decision tree strategy. Moreover, the over-partitioning problem is addressed by the P-Tree methodology, which introduces a threshold ϵ as a stopping condition. If the percentage of instances in a node falls below the predetermined threshold, the expansion of the tree will be halted. The performance of the P-Tree procedure is evaluated on fourteen benchmark data sets with different sizes and contrasted with that of five already existing decision tree methods using a variety of evaluation metrics. The results of the experiments demonstrate that the P-Tree model performs admirably across all of the tested data sets and that it is comparable to the other five decision tree algorithms overall. On the other hand, an ensemble technique called “ensemble P-Tree” offers a reliable remedy to mitigate the instability that is frequently associated with tree-based algorithms. This ensemble method leverages the strengths of the P-Tree approach to enhance predictive performance through collective decision-making. The ensemble P-Tree strategy is comprehensively evaluated by comparing its performance to that of two top-performing ensemble decision tree methodologies. The experimental findings highlight its exceptional performance and competitiveness against other decision tree procedures. Despite the excellent performance of the P-Tree approach, there are still some obstacles that prevent it from handling larger data sets, such as memory restrictions, time complexity, or data complexity. However, parallel computing is effective in resolving this kind of problem. Hence, the MR-P-Tree decision tree technique, a parallel implementation of the P-Tree strategy in the Map-Reduce framework, is further designed. The three parallel procedures MR-SA-S, MR-SP-S, and MR-S-DS for choosing the optimal splitting attributes, choosing the optimal splitting points, and dividing the training data set in parallel, respectively, are the primary basis of the MR-P-Tree methodology. Furthermore, several experimental studies are carried out on ten additional data sets to illustrate the viability of the MR-P-Tree technique and its strong parallel performance.

A hybrid network for fiber orientation distribution reconstruction employing multi-scale information.

Accurate fiber orientation distribution (FOD) is crucial for resolving complex neural fiber structures. However, existing reconstruction methods often fail to integrate both global and local FOD information, as well as the directional information of fixels, which limits reconstruction accuracy. Additionally, these methods overlook the spatial positional relationships between voxels, resulting in extracted features that lack continuity. In regions with signal distortion, many methods also exhibit issues with reconstructionartifacts. This study addresses these challenges by introducing a new neural network calledFusion-Net. Fusion-Net comprises both the FOD reconstruction network and the peak direction estimation network. The FOD reconstruction network efficiently fuses the global and local features of the FOD, providing these features with spatial positional information through a competitive coordinate attention mechanism and a progressive updating mechanism, thus ensuring feature continuity. The peak direction estimation network redefines the task of estimating fixel peak directions as a multi-class classification problem. It uses a direction-aware loss function to supply directional information to the FOD reconstruction network. Additionally, we introduce a larger input scale for Fusion-Net to compensate for local signal distortion by incorporating more globalinformation. Experimental results demonstrate that the rich FOD features contribute to promising performance in Fusion-Net. The network effectively utilizes these features to enhance reconstruction accuracy while incorporating more global information, effectively mitigating the issue of local signaldistortion. This study demonstrates the feasibility of Fusion-Net for reconstructing FOD, providing reliable references for clinicalapplications.

Possibility of Using the Attention Mechanism in Multimodal Recognition of Cardiovascular Diseases

The paper is about studying the possibility of using the attention mechanism in diagnosing various cardiovascular diseases. Biomedical data were presented in different modalities (text, images, and time series). A comparison of the efficiency of 5 transformers based on the attention mechanism (Dosovitsky transformer, compact convolutional trans­former, transformer with external attention, transformer based on tokenization with patch shift and local self-attention, transformer based on multiple deep attention) was carried out with the Exception convolutional neural network, three fully connected neural networks (MLP-Mixer, Fnet, and gMLP), and the YOLO architecture on the problem of multi-class classification (16 classes of dangerous arrhythmias). High efficiency of transformers in diagnosing cardiac diseases was shown. The transformer based on tokenization with patch shift and local self-attention showed the greatest efficiency.

IResNetDM: An interpretable deep learning approach for four types of DNA methylation modification prediction

MotivationAlthough several computational methods for predicting DNA methylation modifications have been developed, two main limitations persist: 1) All of the models are currently confined to binary predictors, which merely determine the presence or absence of DNA methylation modifications and thus prevent comprehensive analyses of the interrelations among varied modification types. Multi-class classification models for RNA modifications have been developed, and a comparable approach for DNA is essential. 2) Few previous studies offer adequate explanations of how models make decisions, instead relying on the extraction and visualization of attention matrices, which have identified few motifs and do not provide sufficient insights into the model decision-making process. ResultIn this study, we introduce the task of DNA methylation modification prediction as a multi-class classification problem for the first time. We present iResNetDM, a deep learning model that integrates Residual Networks (ResNet) with self-attention mechanisms. To the best of our knowledge, iResNetDM is the first model capable of distinguishing between four types of DNA methylation modifications. Our model not only demonstrates good performance across various DNA methylation modifications but can also capture relationships between different types of modifications. We used the integrated gradients technique to enhance the interpretability of the iResNetDM. This method can effectively elucidate the model’s decision-making process, thus enabling the successful identification of multiple motifs. Notably, our model displays remarkable robustness, and can effectively identify unique motifs across different methylation modifications. We also compared the motifs discovered in various modifications and found that some had notable sequence similarities, suggesting that they may be subject to different types of modifications. This finding highlights the potential importance of these motifs in gene regulation.

Exploring online public survey lifestyle datasets with statistical analysis, machine learning and semantic ontology

Lifestyle diseases significantly contribute to the global health burden, with lifestyle factors playing a crucial role in the development of depression. The COVID-19 pandemic has intensified many determinants of depression. This study aimed to identify lifestyle and demographic factors associated with depression symptoms among Indians during the pandemic, focusing on a sample from Kolkata, India. An online public survey was conducted, gathering data from 1,834 participants (with 1,767 retained post-cleaning) over three months via social media and email. The survey consisted of 44 questions and was distributed anonymously to ensure privacy. Data were analyzed using statistical methods and machine learning, with principal component analysis (PCA) and analysis of variance (ANOVA) employed for feature selection. K-means clustering divided the pre-processed dataset into five clusters, and a support vector machine (SVM) with a linear kernel achieved 96% accuracy in a multi-class classification problem. The Local Interpretable Model-agnostic Explanations (LIME) algorithm provided local explanations for the SVM model predictions. Additionally, an OWL (web ontology language) ontology facilitated the semantic representation and reasoning of the survey data. The study highlighted a pipeline for collecting, analyzing, and representing data from online public surveys during the pandemic. The identified factors were correlated with depressive symptoms, illustrating the significant influence of lifestyle and demographic variables on mental health. The online survey method proved advantageous for data collection, visualization, and cost-effectiveness while maintaining anonymity and reducing bias. Challenges included reaching the target population, addressing language barriers, ensuring digital literacy, and mitigating dishonest responses and sampling errors. In conclusion, lifestyle and demographic factors significantly impact depression during the COVID-19 pandemic. The study’s methodology offers valuable insights into addressing mental health challenges through scalable online surveys, aiding in the understanding and mitigation of depression risk factors.

Wavelet-based convolutional neural network for non-intrusive load monitoring of next generation shipboard Power Systems

In this study, a non-intrusive load monitoring (NILM) framework is developed for next generation shipboard power systems (SPS) based on a discrete wavelet transform signal processing and a convolutional neural network (CNN). We have applied the developed NILM method to a four-zone medium voltage direct current (MVDC) SPS to evaluate the effectiveness of the proposed method. Each zone of the MVDC SPS consists of multiple components, such as propulsion load, pulsed load, high ramp rate load, cooling load, and hotel load. The current signals from the main generators are the main inputs to the NILM model. The current signals are first processed through a discrete wavelet transform to create a coefficient vector that reflects the status of all the components in each zone. Then, a multi-class classification problem is formulated and solved using a CNN architecture model to monitor the load statuses in real time. The results of case studies show that the developed NILM model in comparison with benchmark methods can (i) accurately monitor the status of all components with a total accuracy of over 98%, (ii) identify unique pulsed loads with a total accuracy of over 99%, and (iii) sustain the functionality of load monitoring under extreme events such as cyber/physical attacks, load uncertainty, and noisy inputs.

An optimized and intelligent metaverse intrusion detection system based on rough sets

The convergence of the Internet of Things (IoT) and the Metaverse is revolutionizing the digital world by providing immersive, interactive environments as well as new data transmission opportunities. However, this rapid integration raises complex security issues, including an increased risk of unlawful access and data breaches. Strong cybersecurity measures are required to identify and prevent these attacks, preserving the security and confidentiality of users. Finding significant features for recognizing malicious attacks and enhancing the accuracy of network intrusion detection in general, and particularly in the virtual environment, are some of the significant research needs. This paper introduces an intelligent intrusion detection system (IIDS) based on rough set-based electric eel foraging optimization (RSEEFO) in conjunction with the AdaBoost-based classification algorithm. The main objective of this system is to detect and recognize different types of attacks on the interaction and connectivity between the IoT and the metaverse. The proposed IIDS-RSEEFO consists of three phases, which are data pre-processing, multi-IoT attack classification, and evaluation. The main problems associated with the adopted dataset are handled in data pre-processing. Then, in the second phase, a one-versus-all approach is employed along with RSEEFO and AdaBoost to handle the multi-class classification problem. Finally, several evaluation metrics are employed to assess the reliability and robustness of the proposed IIDS-RSEEFO. The proposed IIDS was tested on CIC-IoT-2023 and validated on UNSWNB-15. It achieved high accuracy across all attack types of CIC-IoT-2023, with accuracies of 99.7 %, 100 %, 100 %, 99.8 %, 100 %, 100 %, 99.8 %, and 100 % for benign traffic, DDoS, brute force, spoofing, DoS, Mirai, recon, and web-Based respectively, accompanied by robust sensitivity, F1-Score, Specificity, G-Mean, and crossover-error rate metrics demonstrating its effectiveness in accurately predicting each attack type. Additionally, it obtained high accuracy for all attack types of UNSWNB-15, with accuracies of 96.48 %, 99.12 %, 99.24 %, 93.78 %, 92.55 %, 92.55 %, 92.55 %, 94.73 %, 94.73 %, 98.09 %, 98.39 %, 99.50 %, and 99.95 % for analysis, backdoor, DoS, exploits, fuzzers, generic, normal, reconnaissance, shellcode, and worms, respectively. In addition, the results evaluated that the proposed model is superior compared to the existing intrusion detection systems.

A Novel Multiobjective Genetic Programming Approach to High-Dimensional Data Classification.

The development of data sensing technology has generated a vast amount of high-dimensional data, posing great challenges for machine learning models. Over the past decades, despite demonstrating its effectiveness in data classification, genetic programming (GP) has still encountered three major challenges when dealing with high-dimensional data: 1) solution diversity; 2) multiclass imbalance; and 3) large feature space. In this article, we have developed a problem-specific multiobjective GP framework (PS-MOGP) for handling classification tasks with high-dimensional data. To reduce the large solution space caused by high dimensionality, we incorporate the recursive feature elimination strategy based on mining the archive of evolved GP solutions. A progressive domination Pareto archive evolution strategy (PD-PAES), which optimizes the objectives in a specific order according to their objectives, is proposed to evaluate the GP individuals and maintain a better diversity of solutions. Besides, to address the seriously imbalanced class issue caused by traditional binary decomposition (BD) one versus rest (OVR) for multiclass classification problems, we design a method named BD with a similar positive and negative class size (BD-SPNCS) to generate a set of auxiliary classifiers. Experimental results on benchmark and real-world datasets demonstrate that our proposed PS-MOGP outperforms state-of-the-art traditional and evolutionary classification methods in the context of high-dimensional data classification.

Cost-Sensitive Weighted Contrastive Learning Based on Graph Convolutional Networks for Imbalanced Alzheimer's Disease Staging.

Identifying the progression stages of Alzheimer's disease (AD) can be considered as an imbalanced multi-class classification problem in machine learning. It is challenging due to the class imbalance issue and the heterogeneity of the disease. Recently, graph convolutional networks (GCNs) have been successfully applied in AD classification. However, these works did not handle the class imbalance issue in classification. Besides, they ignore the heterogeneity of the disease. To this end, we propose a novel cost-sensitive weighted contrastive learning method based on graph convolutional networks (CSWCL-GCNs) for imbalanced AD staging using resting-state functional magnetic resonance imaging (rs-fMRI). The proposed method is developed on a multi-view graph constructed by the functional connectivity (FC) and high-order functional connectivity (HOFC) features of the subjects. A novel cost-sensitive weighted contrastive learning procedure is proposed to capture discriminative information from the minority classes, encouraging the samples in the minority class to provide adequate supervision. Considering the heterogeneity of the disease, the weights of the negative pairs are introduced into contrastive learning and they are computed based on the distance to class prototypes, which are automatically learned from the training data. Meanwhile, the cost-sensitive mechanism is further introduced into contrastive learning to handle the class imbalance issue. The proposed CSWCL-GCN is evaluated on 720 subjects (including 184 NCs, 40 SMC patients, 208 EMCI patients, 172 LMCI patients and 116 AD patients) from the ADNI (Alzheimer's Disease Neuroimaging Initiative). Experimental results show that the proposed CSWCL-GCN outperforms state-of-the-art methods on the ADNI database.

Feasibility of Artificial Intelligence Powered Adverse Event Analysis: Using a Large Language Model to Analyze Microwave Ablation Malfunction Data.

Objectives: Determine if a large language model (LLM, GPT-4) can label and consolidate and analyze interventional radiology (IR) microwave ablation device safety event data into meaningful summaries similar to humans. Methods: Microwave ablation safety data from January 1, 2011 to October 31, 2023 were collected and type of failure was categorized by human readers. Using GPT-4 and iterative prompt development, the data were classified. Iterative summarization of the reports was performed using GPT-4 to generate a final summary of the large text corpus. Results: Training (n = 25), validation (n = 639), and test (n = 79) data were split to reflect real-world deployment of an LLM for this task. GPT-4 demonstrated high accuracy in the multiclass classification problem of microwave ablation device data (accuracy [95% CI]: training data 96.0% [79.7, 99.9], validation 86.4% [83.5, 89.0], test 87.3% [78.0, 93.8]). The text content was distilled through GPT-4 and iterative summarization prompts. A final summary was created which reflected the clinically relevant insights from the microwave ablation data relative to human interpretation but had inaccurate event class counts. Conclusion: The LLM emulated the human analysis, suggesting feasibility of using LLMs to process large volumes of IR safety data as a tool for clinicians. It accurately labelled microwave ablation device event data by type of malfunction through few-shot learning. Content distillation was used to analyze a large text corpus (>650 reports) and generate an insightful summary which was like the human interpretation.

Smooth Ordered Weighted Averaging operators

The study demonstrates the application of OWA operators to binary and multiclass classification problems and seeks a way to improve classification accuracy using smoothing methods. OWA operators are used to aggregate class membership probabilities obtained from individual classifiers. Smoothing methods inspired by Newton-Cotes quadratures are applied before the aggregation step to improve the quality of the final results. Moreover, several sets of weights are used for OWA operators, including sets of weights based on the accuracy of individual classifiers. The experiments are conducted on 20 datasets, from which 7 are designed for binary classification and the rest are for multiclass classification. A comparison of the average accuracy for different sets of weights is shown. On the basis of experimental results, smoothing methods that significantly improve classification accuracy are identified.

A dual algorithmic approach to deal with multiclass imbalanced classification problems

Many real-world applications involve multiclass classification problems, and often the data across classes is not evenly distributed. Due to this disproportion, supervised learning models tend to classify instances towards the class with the maximum number of instances, which is a severe issue that needs to be addressed. In multiclass imbalanced data classification, machine learning researchers try to reduce the learning model's bias towards the class with a high sample count. Researchers attempt to reduce this unfairness by either balancing the data before the classifier learns it, modifying the classifier's learning phase to pay more attention to the class with a minimum number of instances, or a combination of both. The existing algorithmic approaches find it difficult to understand the clear boundary between the samples of different classes due to unfair class distribution and overlapping issues. As a result, the minority class recognition rate is poor. A new algorithmic approach is proposed that uses dual decision trees. One is used to create an induced dataset using a PCA based grouping approach and by assigning weights to the data samples followed by another decision tree to learn and predict from the induced dataset. The distinct feature of this algorithmic approach is that it recognizes the data instances without altering their underlying data distribution and is applicable for all categories of multiclass imbalanced datasets. Five multiclass imbalanced datasets from UCI were used to classify the data using our proposed algorithm, and the results revealed that the duo-decision tree approach pays better attention to both the minor and major class samples.

Ordinal Regression With Pinball Loss.

Ordinal regression (OR) aims to solve multiclass classification problems with ordinal classes. Support vector OR (SVOR) is a typical OR algorithm and has been extensively used in OR problems. In this article, based on the characteristics of OR problems, we propose a novel pinball loss function and present an SVOR method with pinball loss (pin-SVOR). Pin-SVOR is fundamentally different from traditional SVOR with hinge loss. Traditional SVOR employs the hinge loss function, and the classifier is determined by only a few data points near the class boundary, called support vectors, which may lead to a noise sensitive and re-sampling unstable classifier. Distinctively, pin-SVOR employs the pinball loss function. It attaches an extra penalty to correctly classified data that lies inside the class, such that all the training data is involved in deciding the classifier. The data near the middle of each class has a small penalty, and that near the class boundary has a large penalty. Thus, the training data tend to lie near the middle of each class instead of on the class boundary, which leads to scatter minimization in the middle of each class and noise insensitivity. The experimental results show that pin-SVOR has better classification performance than state-of-the-art OR methods.

Explainable dating of greek papyri images

Greek literary papyri, which are unique witnesses of antique literature, do not usually bear a date. They are thus currently dated based on palaeographical methods, with broad approximations which often span more than a century. We created a dataset of 242 images of papyri written in “bookhand” scripts whose date can be securely assigned, and we used it to train algorithms for the task of dating, showing its challenging nature. To address data scarcity, we extended our dataset by segmenting each image into its respective text lines. By using the line-based version of our dataset, we trained a Convolutional Neural Network, equipped with a fragmentation-based augmentation strategy, and we achieved a mean absolute error of 54 years. The results improve further when the task is cast as a multi-class classification problem, predicting the century. Using our network, we computed precise date estimations for papyri whose date is disputed or vaguely defined, employing explainability to understand dating-driving features.

Neural network learning and Quantum Gravity

The landscape of low-energy effective field theories stemming from string theory is too vast for a systematic exploration. However, the meadows of the string landscape may be fertile ground for the application of machine learning techniques. Employing neural network learning may allow for inferring novel, undiscovered properties that consistent theories in the landscape should possess, or checking conjectural statements about alleged characteristics thereof. The aim of this work is to describe to what extent the string landscape can be explored with neural network-based learning. Our analysis is motivated by recent studies that show that the string landscape is characterized by finiteness properties, emerging from its underlying tame, o-minimal structures. Indeed, employing these results, we illustrate that any low-energy effective theory of string theory is endowed with certain statistical learnability properties. Consequently, several local learning problems therein formulated, including interpolations and multi-class classification problems, can be concretely addressed with machine learning, delivering results with sufficiently high accuracy.

Safe dynamic sparse training of modified RBF networks for joint feature selection and classification

Developing an efficient embedded feature selection method for both binary and multiclass classification problems is a fundamental topic to be further studied. In this paper, the conventional radial basis function (RBF) network is modified for joint feature selection and classification. By utilizing the anisotropic Gaussian basis function, an individual weight parameter is assigned to each feature for feature weighting. The L1-regularized loss function is designed by combining the softmax cross entropy loss with L1 regularization terms imposing on feature and output weights. A specialized active-set limited-memory projected Quasi-Newton (SAL-PQN) algorithm is proposed to minimize the L1-regularized loss function, which can optimize all the adjustable parameters and zero out the redundant feature and output weights simultaneously. Furthermore, a safe dynamic pruning strategy is integrated into the SAL-PQN algorithm for dynamic sparse training of the modified RBF network, which continuously excludes the dynamically zeroed-out feature weights from the training process. It is theoretically assured of yielding the same gradient updates as SAL-PQN, thus the relevant computation of the zeroed-out feature weights in the forward and backward propagations can be safely eliminated. The experimental results demonstrate the effectiveness and efficiency of the proposed methods for joint feature selection and classification, and illustrate the practical utility of the inherent feature weighting capability.

An automated ensemble approach using Harris Hawk optimization for visually evoked EEG signal classification.

Steady-state visually evoked potential is one of the active explorations in the brain-computer interface research. Electroencephalogram based brain computer interface studies have been widely applied to perceive solutions for real-world problems in the healthcare domain. The classification of externally bestowed visual stimuli of different frequencies on a human was experimented to identify the need of paralytic people. Although many classifiers are at the fingertip of machine learning technology, recent research has proven that ensemble learning is more efficacious than individual classifiers. Despite its efficiency, ensemble learning technology exhibits certain drawbacks like taking more time on selecting the optimal classifier subset. This research article utilizes the Harris Hawk Optimization algorithm to select the best classifier subset from the given set of classifiers. The objective of the research is to develop an efficient multi-classifier model for electroencephalogram signal classification. The proposed model utilizes the Boruta Feature Selection algorithm to select the prominent features for classification. Thus selected prominent features are fed into the multi-classifier subset which has been generated by the Harris Hawk Optimization algorithm. The results of the multi-classifier ensemble model are aggregated using Stacking, Bagging, Boosting, and Voting. The proposed model is evaluated against the acquired dataset and produces a promising accuracy of 96.1%, 98.7%, 91.91%, and 99.01% with the ensemble techniques respectively. The proposed model is also validated with other performance metrics such as sensitivity, specificity, and F1-Score. The experimental results show that the proposed model proves its supremacy in segregating the multi-class classification problem with high accuracy.

ENHANCING MULTI-CLASS TEXT CLASSIFICATION WITH APRIORI-BASED FEATURE SELECTION

: In the field of Natural Language Processing, selecting the right features is crucial for reducing unnecessary model complexity, speeding up training, and improving the ability to generalize. However, the multi-class text classification problem makes it challenging for models to generalize well, which complicates feature selection. This paper investigates how feature selection impacts model performance for multi-class text classification, using a dataset of projects completed by TÜBİTAK TEYDEB between 2009 and 2022. The study employs LSTM, a deep learning method, to classify the projects into nine different industries based on various attributes. The paper proposes a new feature selection approach based on the Apriori algorithm, which reduces the number of attribute combinations considered and makes model training more efficient. Model performance is evaluated using metrics like accuracy, loss, validation scores, and test scores. The key findings are that feature selection significantly affects model performance, and different feature sets have varying impacts on performance.

Dual-stage deep learning for sangac optical fiber sensing multi-event detection and localization

This paper introduces a dual-stage deep learning structure for multi-event detection and localization (DDMDL) for a loop-based Sagnac interferometer optical fiber sensing system to overcome the difficulties of detecting and localizing multiple events. The DDMDL approach comprises of a combination of (i) a detection stage made of a multi-class classification model that quantifies the events, and (ii) a localization stage made of separate deep-learning regression models tasked with precisely localizing the events. By addressing the multi-event localization as a combined multi-class classification and multi-regression problem, our approach enables more accurate prediction of event locations. We evaluate our DDMDL model in a broad range of test scenarios across three signal-to-noise ratio (SNR) levels: low, medium, and high, all outside the model’s training dataset. The model had high detection (i.e., classification) accuracy in simulation, with rates of 90%–99% for single-event scenarios, 94%–98% for two-event scenarios, and 82%–99% for three-event scenarios at the three SNR levels. The precision of localization (i.e., regression) was evaluated by Mean Absolute Error (MAE). The results showed that single-event scenarios could be localized within a range of 35–50 m, two-event scenarios within a range of 82–106 m, and three-event scenarios within a range of 131–151 m. These findings were consistent across different SNR levels, ranging from low to high, over a 50 km sensing fiber length. Experimental validations confirmed the DDMDL model’s practical applicability, with detection accuracy of 80% in single-event scenarios and localization accuracy with MAEs ranging between 32 and 65 m. For two-event scenarios, the model achieved an 87% success rate, with MAEs ranging between 122 m and 204 m, emphasizing the model’s potential effectiveness in different applications.

A Study on the Robustness and Stability of Explainable Deep Learning in an Imbalanced Setting: The Exploration of the Conformational Space of G Protein-Coupled Receptors.

G-protein coupled receptors (GPCRs) are transmembrane proteins that transmit signals from the extracellular environment to the inside of the cells. Their ability to adopt various conformational states, which influence their function, makes them crucial in pharmacoproteomic studies. While many drugs target specific GPCR states to exert their effects-thereby regulating the protein's activity-unraveling the activation pathway remains challenging due to the multitude of intermediate transformations occurring throughout this process, and intrinsically influencing the dynamics of the receptors. In this context, computational modeling, particularly molecular dynamics (MD) simulations, may offer valuable insights into the dynamics and energetics of GPCR transformations, especially when combined with machine learning (ML) methods and techniques for achieving model interpretability for knowledge generation. The current study builds upon previous work in which the layer relevance propagation (LRP) technique was employed to interpret the predictions in a multi-class classification problem concerning the conformational states of the β2-adrenergic (β2AR) receptor from MD simulations. Here, we address the challenges posed by class imbalance and extend previous analyses by evaluating the robustness and stability of deep learning (DL)-based predictions under different imbalance mitigation techniques. By meticulously evaluating explainability and imbalance strategies, we aim to produce reliable and robust insights.