Popular Metrics Research Articles

YOLOv7‐XAI: Multi‐Class Skin Lesion Diagnosis Using Explainable AI With Fair Decision Making

ABSTRACTSkin cancer, a prevalent and potentially life‐threatening condition, demands accurate and timely detection for effective intervention. It is an uncontrolled growth of abnormal cells in the human body. Studies are underway to determine if a skin lesion is benign (non‐cancerous) or malignant (cancerous), but the biggest challenge for a doctor is determining the type of skin cancer. As a result, determining the type of tumour is crucial for the right course of treatment. In this study, we introduce a groundbreaking approach to multi‐class skin cancer detection by harnessing the power of Explainable Artificial Intelligence (XAI) in conjunction with a customised You Only Look Once (YOLOv7) architecture. Our research focuses on enhancing the YOLOv7 framework to accurately discern 8 different skin cancer classes, including melanoma, basal cell carcinoma, and squamous cell carcinoma. The YOLOv7 model is the robust backbone, enriched with features tailored for precise multi‐class classification. Concurrently, integrating XAI elements, Local Interpretable Modal‐agnostic Explanations (LIME) and Shapley Additive Explanations (SHAP) ensures transparent decision‐making processes, enabling healthcare professionals to interpret and trust the model's predictions. This innovative synergy between YOLOv7 and XAI heralds a new era in interpretability, resulting in high‐performance skin cancer diagnostics. The obtained results are 96.8%, 94.2%, 95.6%, and 95.8%, evaluated with popular quantitative metrics such as accuracy, precision, recall, and F1 score, respectively.

An exploration of the impact of phylogenetic tree structure on NTI and βNTI estimates of community assembly

The nearest taxon index (NTI) and beta nearest taxon index (βNTI) are popular metrics used to infer community assembly in a variety of ecosystems. However, the dependence these metrics have on the underlying phylogenetic tree is often overlooked and can potentially lead to erroneous conclusions. In this paper, we investigated how misclassified sequences in the phylogenetic tree and reduction of the metacommunity richness affect estimates of the assembly processes. We estimated assembly processes in two datasets investigating fish-microbiota interactions. We found that non-bacterial misclassified sequences can drastically change the interpretation of assembly processes. NTI consistently increased after misclassified sequences were removed, which changed the ecological interpretation for 82.8% of the communities in one of the datasets. Similarly, βNTI was strongly affected by misclassified sequences in the phylogenetic tree, changing the ecological interpretation for 44.7% (Dataset 1) and 6.3% (Dataset 2) of the community comparisons. Lower metacommunity richness generally resulted in decreased NTI and increased βNTI, but the effect sizes were small and generally did not result in a change in ecological interpretation. Based on these findings, we recommend: (1) constructing and visually inspecting a phylogenetic tree for 16 S rRNA gene amplicon datasets to identify and remove misclassified sequences, and (2) evaluating the robustness of community assembly conclusions by performing analyses with varying metacommunity richness and clearly specifying the metacommunity used for the null model approach.

MG-Net: A fetal brain tissue segmentation method based on multiscale feature fusion and graph convolution attention mechanisms

Background and Objective:Fetal brain tissue segmentation provides foundational support for comprehensively understanding the neurodevelopment of normal and congenital disease-affected fetuses. Manual labeling is very time-consuming, and automated segmentation methods can greatly improve the efficiency of doctors. At the same time, fetal brain tissue undergoes various changes throughout the pregnancy, leading to a continuous change in tissue contrast, which greatly increases the difficulty of training segmentation methods. This study aims to develop an automated segmentation model that can efficiently and accurately segment fetal brain tissue, improving the workflow for medical professionals. Methods:We propose a novel deep learning-based segmentation model that incorporates three innovative components: Firstly, a new Dual Dilated Attention Block (DDAB) is proposed in the encoder part to enhance the feature extraction of local spatial and structural contextual information. Secondly, a Multi-scale Deformable Transformer (MSDT) is integrated into the bottleneck to improve the feature extraction of global information on local spatial and structural contextual information. Thirdly, we use a novel block based on Graph Convolution Attention (GCAB) in the decoder, which effectively enhances the features at the decoder.The code is available at https://github.com/unicoco7/MG-Net/. Results:We trained and tested on the FeTA 2021 and FeTA 2022 datasets, and evaluated using seven popular metrics, including Dice, IoU, MAE, BoundaryF, PRE, SEN, and SPE. Compared to the current state-of-the-art 3D segmentation models such as nnFormer, SwinUNETR, and 3DUX-net, our proposed method has surpassed all of them in metrics like Dice, IoU, and MAE. Specifically, on the FeTA 2021 dataset, our model achieved a Dice of 0.8666, an IoU of 0.7646, and an MAE of 0.0027; on the FeTA 2022 dataset, it achieved a Dice of 0.8552, an IoU of 0.7470, and an MAE of 0.0005. Conclusion:In this paper, we propose a model for three-dimensional fetal brain tissue segmentation based on multi-scale feature fusion and graph convolution attention mechanism, and conduct experimental evaluation on the FeTA 2021 and FeTA 2022 datasets. Understanding the boundaries of fetal brain tissue is crucial for doctors’ diagnosis, so the proposed model is expected to improve the speed and accuracy of doctors’ diagnoses.

Moderators of curiosity and information seeking in younger and older adults.

The present study examined age differences in the influence of informational value cues on curiosity and information seeking. In two experiments, younger and older adults (total N = 514) rated their curiosity about content before having the opportunity to seek out more information. Experiment 1 examined the impact of social value on curiosity and information seeking about trivia. Online popularity metrics served as social value cues. Metric visibility increased engagement with high-popularity information for older adults, whereas it decreased engagement with low-popularity information for younger adults. Experiment 2 examined the impact of practical value on curiosity and information seeking about science facts. Personal and collective practical value were highlighted by linking the information to the domains of medicine and the environment, respectively. Patterns of curiosity and information seeking revealed greater sensitivity to collective practical value in older than younger adults. In both experiments, the relationship between curiosity and information seeking was stronger in older adults than in younger adults. Overall, these findings suggest that age differences in motivational priorities may lead to age differences in curiosity and information seeking. In addition to highlighting strategies for fostering curiosity in older learners, these findings may also inform digital literacy interventions aimed at reducing engagement with clickbait and misinformation. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Synthetic oversampling with Mahalanobis distance and local information for highly imbalanced class-overlapped data

Minority oversampling is currently one of the most popular and effective methods for handling imbalanced data. However, oversampling that relies on the observations of the minority class to generate new samples is not applicable in the scenario of imbalanced data with extremely scarce minority samples, because the strongly underrepresented minority class does not contain enough information to support the oversampling process. Since some recent studies have exhibited the effectiveness of using majority information to bootstrap oversampling, the neglect of class overlap in the sampling process would increase the overlapping degree and complicate the decision boundary. To this end, this paper proposes a Mahalanobis distance and Local information based OverSampling (MLOS) for highly imbalanced class-overlapped data. MLOS first employs the majority density to guide the sample synthesis, with Mahalanobis distance to extract the majority probability contour. Then for each minority seed sample, to avoid the generation of overlapping samples, MLOS constrain the synthetic process by finding the auxiliary sample (in its 5 nearest neighbors) with similar probability density value to the seed. Finally, MLOS uses a pair-wise data cleaning process to improve the visibility of the decision boundary according to the probability density of synthetic samples. Comparative experiments conducted on 16 highly imbalanced class-overlapped datasets, using 17 different methods, demonstrates the superiority of our proposed method in terms of three popular evaluation metrics AUC, G-mean and Recall for imbalance classification. The source code of MLOS is available at https://github.com/ytyancp/MLOS.

Accuracy of Relative Valuation Methods in Predicting the Offer Price of Indonesian IPO Stocks

The increasing Initial Public Offerings (IPO) trend in Indonesia has created many investment opportunities for institutional and retail investors. However, multiple valuation methods often result in varying IPO price predictions, leading to uncertainty and confusion among investors. This study aims to assess the effectiveness of relative valuation methods in predicting IPO offer prices, focusing on widely adopted methods due to their simplicity and accessibility. The research evaluates three popular relative valuation metrics: the price-to-book (P/B) ratio, the price-to-earnings (P/E) ratio, and the price-to-sales (P/S) ratio. By analyzing companies that went public in Indonesia in 2023 and utilizing average financial ratios from the preceding year, the study seeks to determine which method provides the most accurate price predictions. The findings indicate that the P/B ratio is the most reliable method, achieving an average absolute error of 51.1%. This suggests that the P/B method minimizes valuation errors more effectively than P/E and P/S ratios. Moreover, regression models' P/B ratio demonstrates the highest adjusted R² value of 0.781, further validating its predictive accuracy. The implications of this study are significant for investors and financial analysts who rely on relative valuation methods to make informed decisions about IPO investments. By highlighting the P/B ratio's superior performance in predicting IPO offer prices, this research contributes to the ongoing discourse on valuation methodologies and their practical application in emerging markets like Indonesia.

Forecasting Stock Market Performance: An Ensemble Learning-Based Approach

There is tremendous value in the ability to predict stock market trends and outcomes. The public sentiment surrounding a stock is unquestionably a vital factor contributing to the rise or fall of a stock price. This paper aims to detail how data from public sentiment can be integrated into traditional stock analyses and how these analyses can then be used to make predictions of stock price trends. Headlines from seven news publications and conversations from Yahoo! Finance’s conversations forum were processed by the Valence Aware Dictionary and sEntiment Reasoner (VADER) natural language processing package to determine numerical polarities that represent a positive, negative, or neutral public sentiment around a stock ticker. The resulting polarities were paired with popular stock-table metrics to create a dataset for a Logistic Regression machine learning model. The model was trained on approximately 4400 major stocks to determine a binary “Buy” (1) or “Not Buy” (0) recommendation for each stock. We present approaches for a variety of machine learning models – Logistic Regression, Random Forest Classifiers, and Extreme Gradient Boosting models and do a statistical comparison on the balanced accuracies of each machine learning model. This paper aims to detail our understanding of how one can leverage the public sentiment of a stock to gauge its future value.

Diverse representation-guided graph learning for multi-view metric clustering

Multi-view graph clustering has garnered tremendous interest for its capability to effectively segregate data by harnessing information from multiple graphs representing distinct views. Despite the advances, conventional methods commonly construct similarity graphs straightway from raw features, leading to suboptimal outcomes due to noise or outliers. To address this, latent representation-based graph clustering has emerged. However, it often hypothesizes that multiple views share a fixed-dimensional coefficient matrix, potentially resulting in useful information loss and limited representation capabilities. Additionally, many methods exploit Euclidean distance as a similarity metric, which may inaccurately measure linear relationships between samples. To tackle these challenges, we develop a novel diverse representation-guided graph learning for multi-view metric clustering (DRGMMC). Concretely, raw sample matrix from each view is first projected into diverse latent spaces to capture comprehensive knowledge. Subsequently, a popular metric is leveraged to adaptively learn similarity graphs with linearity-aware based on attained coefficient matrices. Furthermore, a self-weighted fusion strategy and Laplacian rank constraint are introduced to output clustering results directly. Consequently, our model merges diverse representation learning, metric learning, consensus graph learning, and data clustering into a joint model, reinforcing each other for holistic optimization. Substantial experimental findings substantiate that DRGMMC outperforms most advanced graph clustering techniques.

Modelling lethality and teratogenicity of zebrafish (Danio rerio) due to β-lactam antibiotics employing the QSTR approach

ABSTRACT Nowadays, β-lactam antibiotics are one of the most consumed OTC (over-the-counter) medicines in the world. Its frequent use against several infectious diseases leads to the development of antibiotic resistance. Another unavoidable risk factor of β-lactam antibiotics is environmental toxicity. Numerous terrestrial as well as aquatic species have suffered due to the excessive use of these pharmaceuticals. In this present study, we have performed a toxicity assessment employing a novel in silico technique like quantitative structure–toxicity relationships (QSTRs) to explore toxicity against zebrafish (Danio rerio). We have developed single as well as inter-endpoint QSTR models for the β-lactam compounds to explore important structural attributes responsible for their toxicity, employing median lethal (LC50) and median teratogenic concentration (TC50) as the endpoints. We have shown how an inter-endpoint model can extrapolate unavailable endpoint values with the help of other available endpoint values. To verify the models’ robustness, predictivity, and goodness-of-fit, several universally popular metrics for both internal and external validation were extensively employed in model validation (single endpoint models: r 2 = 0.631 − 0.75, Q 2 F1 = 0.607 − 0.684; inter-endpoint models: r 2 = 0.768 − 0.84, Q 2 F1 = 0.678 − 0.76). Again, these models were engaged in the prediction of these two responses for a true external set of β-lactam molecules without response values to prove the reproducibility of these models.

Leveraging Machine Learning and Patient Reviews for Developing a Drug Recommendation System to Reduce Medical Errors

Background In the rapidly evolving pharmaceutical industry, drug efficacy and safety stand as critical concerns. The vast accumulation of data, including customer feedback, drug popularity, and usage details, offers a rich resource for improving healthcare outcomes. Aims The primary aim of this study is to harness machine learning and Natural Language Processing (NLP) techniques to sift through extensive pharmaceutical data, identifying the most effective drugs for various conditions and uncovering patterns that could guide better decision-making in drug efficacy and safety. Objective This research seeks to construct a sophisticated model capable of analyzing diverse data points to pinpoint the most efficacious drugs for specific health conditions, thereby providing pharmaceutical companies with data-driven insights to optimize drug safety and effectiveness. Methods Employing a blend of Natural Language Processing (NLP) and machine learning strategies, the study analyzes a comprehensive dataset featuring customer reviews, drug popularity metrics, usage information, and other relevant data collected over an extended period. This methodological approach aims to reveal latent trends and patterns that are crucial for assessing drug performance. Results The developed model adeptly identifies leading medications for various conditions, elucidating efficacy and safety profiles derived from patient reviews and drug utilization trends. These findings furnish pharmaceutical companies with actionable intelligence for enhancing drug development and patient care strategies. Conclusion The integration of machine learning and NLP for the analysis of vast drug-related datasets presents a powerful method for advancing drug efficacy and safety. The insights yielded by the proposed model significantly empower the decision-making processes of the pharmaceutical industry, ultimately fostering improved health outcomes for patients.

Is YouTube a Sufficient and Reliable Source to Inform Patients About Cardiac Rehabilitation?: A Cross-sectional Study.

There is a growing concern surrounding the utility of medical content on social media. In this study, the popularity metrics and content quality of cardiac rehabilitation (CR) videos on YouTube regarding patient education were examined. Using the search key word "cardiac rehabilitation," we analyzed the 50 most relevant videos. Our video popularity analytics encompassed viewing rate, such as ratio, number of comments, and the video power index (VPI). We assessed content quality using the Global Quality Scale (GQS), the modified DISCERN questionnaire, Journal of the American Medical Association ( JAMA ) benchmark criteria, Patient Education Materials Assessment Tool for Audio/Visual Materials (PEMAT-A/V), and a novel tool, the Cardiac Rehabilitation Specific Scale (CRSS). Notably, 78% of the videos were uploaded by medical organizations. The average viewing rate was 4.6 views per day. There were positive correlations between the scores from different content quality scales. Median scores for the GQS, the modified DISCERN questionnaire, JAMA benchmark criteria, and the CRSS were 3, 3.5, 2, and 5, respectively. Mean PEMAT-A/V scores were 60.4% for understandability and 38.3% for actionability. Videos published by entities other than medical centers predicted lower CRSS and GQS scores. High JAMA benchmark criteria scores were negative predictors of VPI, view rate, and number of comments. Our findings suggest that CR-related videos on YouTube are characterized by low popularity, average content quality and understandability, but a lack of reliability and actionability. To ensure individuals seek accurate CR information on social media platforms, we recommend directing them to videos uploaded by medical centers.

Generalized multikernel correntropy based broad learning system for robust regression

As an emerging learning method belonging to the family of neural networks, the broad learning system (BLS) has been recently proved to be effective and efficient to perform regression tasks in various scenarios. However, if data are contaminated by some outliers or other more complex non-Gaussian noises, the learning performance of BLS may be severely compromised, due to its dependence on the conventional mean square error criterion. To enhance the robustness of BLS to deal with contaminated data, a new similarity measure termed generalized multikernel correntropy (GMKC) is proposed in this paper, and some important properties of this measure are investigated. On the basis of GMKC, a general BLS variant called GMKC-based BLS (GMKC-BLS), is subsequently developed to perform regression tasks with contaminated data. Since GMKC with its unique design actually builds a unified framework for many robust and popular metrics, GMKC-BLS is expected to be with excellent robustness and adaptability, and provides a competitive solution to the regression problems with contaminated data. Meanwhile, GMKC could be integrated with other neural network-based methods to further enhance their robustness. Experimental results on different regression datasets demonstrate the performance superiority of GMKC-BLS compared to the standard BLS and its robust variants.

Improving ROUGE‐1 by 6%: A novel multilingual transformer for abstractive news summarization

SummaryNatural language processing (NLP) has undergone a significant transformation, evolving from manually crafted rules to powerful deep learning techniques such as transformers. These advancements have revolutionized various domains including summarization, question answering, and more. Statistical models like hidden Markov models (HMMs) and supervised learning have played crucial roles in laying the foundation for this progress. Recent breakthroughs in transfer learning and the emergence of large‐scale models like BERT and GPT have further pushed the boundaries of NLP research. However, news summarization remains a challenging task in NLP, often resulting in factual inaccuracies or the loss of the article's essence. In this study, we propose a novel approach to news summarization utilizing a fine‐tuned Transformer architecture pre‐trained on Google's mt‐small tokenizer. Our model demonstrates significant performance improvements over previous methods on the Inshorts English News dataset, achieving a 6% enhancement in the ROUGE‐1 score and reducing training loss by 50%. This breakthrough facilitates the generation of reliable and concise news summaries, thereby enhancing information accessibility and user experience. Additionally, we conduct a comprehensive evaluation of our model's performance using popular metrics such as ROUGE scores, with our proposed model achieving ROUGE‐1: 54.6130, ROUGE‐2: 31.1543, ROUGE‐L: 50.7709, and ROUGE‐LSum: 50.7907. Furthermore, we observe a substantial reduction in training and validation losses, underscoring the effectiveness of our proposed approach.

Introducing Grid WAR: rethinking WAR for starting pitchers

Abstract The baseball statistic “Wins Above Replacement” (WAR) has emerged as one of the most popular evaluation metrics. But it is not readily observed and tabulated; WAR is an estimate of a parameter in a vaguely defined model with all its attendant assumptions. Industry-standard models of WAR for starting pitchers from FanGraphs and Baseball Reference all assume that season-long averages are sufficient statistics for a pitcher’s performance. This provides an invalid mathematical foundation for many reasons, especially because WAR should not be linear with respect to any counting statistic. To repair this defect, as well as many others, we devise a new measure, Grid WAR, which accurately estimates a starting pitcher’s WAR on a per-game basis. The convexity of Grid WAR diminishes the impact of “blow-up” games and up-weights exceptional games, raising the valuation of pitchers like Sandy Koufax, Whitey Ford, and Catfish Hunter who exhibit fundamental game-by-game variance. Although Grid WAR is designed to accurately measure historical performance, it has predictive value insofar as a pitcher’s Grid WAR is better than Fangraphs’ FIP WAR at predicting future performance. Finally, at https://gridwar.xyz we host a Shiny app which displays the Grid WAR results of each MLB game since 1952, including career, season, and game level results, which updates automatically every morning.

Measuring the centrality of nodes in networks based on the interstellar model

Measuring the centrality of nodes in a network is a vital and significant task in analyzing the influence of nodes and information dissemination. Existing methods measure the centrality of nodes mainly from local, global and semi-global network perspectives. Currently, semi-global centrality metrics are valued due to the fact that they can achieve high accuracy close to that of global centrality metrics with a slight increase in time complexity over local centrality metrics. In this paper, a novel semi-global centrality metric based on the interstellar model is proposed. First, an interstellar model of information dissemination is constructed and a variable velocity is set for the information after it leaves the source node, while velocity at which the information arrives at the next node determines whether the information can be further forwarded. Then, the attraction slingshot effect is considered to express the facilitation provided by the nodes that forward the information. The centrality of a node can be measured by the number of all nodes, including itself, that can forward the information sent by that node. Experiments with some popular centrality metrics are conducted on some real-world networks, and the results show that the proposed centrality metric has better performance without significant increase in time complexity, and also provides a plausible explanation for information forwarding.

TIST-Net: style transfer in dynamic contrast enhanced MRI using spatial and temporal information

Objective. Training deep learning models for image registration or segmentation of dynamic contrast enhanced (DCE) MRI data is challenging. This is mainly due to the wide variations in contrast enhancement within and between patients. To train a model effectively, a large dataset is needed, but acquiring it is expensive and time consuming. Instead, style transfer can be used to generate new images from existing images. In this study, our objective is to develop a style transfer method that incorporates spatio-temporal information to either add or remove contrast enhancement from an existing image. Approach. We propose a temporal image-to-image style transfer network (TIST-Net), consisting of an auto-encoder combined with convolutional long short-term memory networks. This enables disentanglement of the content and style latent spaces of the time series data, using spatio-temporal information to learn and predict key structures. To generate new images, we use deformable and adaptive convolutions which allow fine grained control over the combination of the content and style latent spaces. We evaluate our method, using popular metrics and a previously proposed contrast weighted structural similarity index measure. We also perform a clinical evaluation, where experts are asked to rank images generated by multiple methods. Main Results. Our model achieves state-of-the-art performance on three datasets (kidney, prostate and uterus) achieving an SSIM of 0.91 ± 0.03, 0.73 ± 0.04, 0.88 ± 0.04 respectively when performing style transfer between a non-enhanced image and a contrast-enhanced image. Similarly, SSIM results for style transfer from a contrast-enhanced image to a non-enhanced image were 0.89 ± 0.03, 0.82 ± 0.03, 0.87 ± 0.03. In the clinical evaluation, our method was ranked consistently higher than other approaches. Significance. TIST-Net can be used to generate new DCE-MRI data from existing images. In future, this may improve models for tasks such as image registration or segmentation by allowing small training datasets to be expanded.

A personalized product recommendation model in e-commerce based on retrieval strategy

In recent years, online shopping is one of the routine parts in people’s life. It is convenient and takes less effort to purchase it. Regarding the increasing revolution of e-commerce businesses, recommendation engine plays a crucial role in them. Recommendation engines are very popular and easy to implement to their platform nowadays. Due to the extremely high competition of e-commerce businesses, the operation needs to integrate the recommender wisely. This study presents a comprehensive approach to improving user experience and engagement on e-commerce platforms through the implementation of an implicit personalized product recommendation engine. Collaborating with the H&M Group, the research combines the strength of each recommending algorithms which are collaborative filtering, popularity, and Bayesian personalized ranking to develop a robust recommendation system. By leveraging a retrieval strategy that combines multiple algorithmic techniques and evaluating candidates using machine learning models which comprise LightGBM and Deep Neural Network, the study achieves promising results. The authors utilize two popular technical metrics to evaluate their models which are mean average precision at K candidates (MAP@K) and mean average recall at K candidates (MAR@K). The empirical result indicates that the LightGBM model has remarkable performance than Deep Neural Network model, which are 0.06 versus 0.02 respectively in MAP@K and 0.03 versus 0.01 respectively in MAR@K when both recommending ways is at 50 items. Overall, this research contributes a novel framework that addresses the challenges of analyzing large-scale data, cold-start problems, and personalization, thereby enhancing the user experience, and driving sales on e-commerce platforms.

Predicting human effort needed to correct auto-segmentations.

Medical image auto-segmentation techniques are basic and critical for numerous image-based analysis applications that play an important role in developing advanced and personalized medicine. Compared with manual segmentations, auto-segmentations are expected to contribute to a more efficient clinical routine and workflow by requiring fewer human interventions or revisions to auto-segmentations. However, current auto-segmentation methods are usually developed with the help of some popular segmentation metrics that do not directly consider human correction behavior. Dice Coefficient (DC) focuses on the truly-segmented areas, while Hausdorff Distance (HD) only measures the maximal distance between the auto-segmentation boundary with the ground truth boundary. Boundary length-based metrics such as surface DC (surDC) and Added Path Length (APL) try to distinguish truly-predicted boundary pixels and wrong ones. It is uncertain if these metrics can reliably indicate the required manual mending effort for application in segmentation research. Therefore, in this paper, the potential use of the above four metrics, as well as a novel metric called Mendability Index (MI), to predict the human correction effort is studied with linear and support vector regression models. 265 3D computed tomography (CT) samples for 3 objects of interest from 3 institutions with corresponding auto-segmentations and ground truth segmentations are utilized to train and test the prediction models. The five-fold cross-validation experiments demonstrate that meaningful human effort prediction can be achieved using segmentation metrics with varying prediction errors for different objects. The improved variant of MI, called MIhd, generally shows the best prediction performance, suggesting its potential to indicate reliably the clinical value of auto-segmentations.

DGCLUSTER: A Neural Framework for Attributed Graph Clustering via Modularity Maximization

Graph clustering is a fundamental and challenging task in the field of graph mining where the objective is to group the nodes into clusters taking into consideration the topology of the graph. It has several applications in diverse domains spanning social network analysis, recommender systems, computer vision, and bioinformatics. In this work, we propose a novel method, DGCluster, which primarily optimizes the modularity objective using graph neural networks and scales linearly with the graph size. Our method does not require the number of clusters to be specified as a part of the input and can also leverage the availability of auxiliary node level information. We extensively test DGCluster on several real-world datasets of varying sizes, across multiple popular cluster quality metrics. Our approach consistently outperforms the state-of-the-art methods, demonstrating significant performance gains in almost all settings.

Error-robust multi-view subspace clustering with nonconvex low-rank tensor approximation and hyper-Laplacian graph embedding

Tensor based subspace clustering, a method aimed at partitioning multi-view data into distinct clusters through the tensor low-rank representation, has gained widespread popularity. Despite the effectiveness of tensor singular value decomposition (t-SVD) based nuclear norm in extracting high-dimensional information from multiple views, certain limitations persist. Firstly, while the commonly used l2,1 norm enhances clustering robustness, it remains susceptible to the impact of high-intensity noises. Secondly, the t-SVD based nuclear norm provides a biased estimation of tensor rank. Thirdly, existing methods either fail to preserve manifold structures from the original space or only extract binary relationships between data points. To address these challenges and enhance clustering performance in both high-intensity noisy and real-world conditions, an error-robust multi-view subspace clustering with nonconvex low-rank tensor approximation and hyper-Laplacian graph embedding (EMSC-NLTHG) method is proposed. Specifically, Cauchy loss function which reduces sensitivity to larger noises and outliers is introduced. Based on the Cauchy loss function, we develop the Cauchy pseudo norm to represent the reconstruction error in clustering to enhance the robustness. Moreover, rather than uniformly weighting all singular values in the t-SVD nuclear norm, a nonconvex tensor nuclear norm is adopted to add weights adaptively and approximate the true tensor rank. Additionally, hypergraph is embedded in the representation matrices to preserve local manifold structures and extract the complex multivariate relationships between data points. Extensive experiments demonstrate that the proposed EMSC-NLTHG method consistently outperforms state-of-the-art techniques by approximately 18% to 40% on datasets corrupted by high-intensity noises and around 1% to 27% on eight real-world datasets across six popular evaluation metrics.