Project Dataset Research Articles

Efficient and Accurate Zero-Day Electricity Theft Detection from Smart Meter Sensor Data Using Prototype and Ensemble Learning

Electricity theft remains a pressing challenge in modern smart grid systems, leading to significant economic losses and compromised grid stability. This paper presents a sensor-driven framework for electricity theft detection that leverages data collected from smart meter sensors, key components in smart grid monitoring infrastructure. The proposed approach combines prototype learning and meta-level ensemble learning to develop a scalable and accurate detection model, capable of identifying zero-day attacks that are not present in the training data. Smart meter data is compressed using Principal Component Analysis (PCA) and K-means clustering to extract representative consumption patterns, i.e., prototypes, achieving a 92% reduction in dataset size while preserving critical anomaly-relevant features. These prototypes are then used to train base-level one-class classifiers, specifically the One-Class Support Vector Machine (OCSVM) and the Gaussian Mixture Model (GMM). The outputs of these classifiers are normalized and fused in a meta-OCSVM layer, which learns decision boundaries in the transformed score space. Experimental results using the Irish CER Smart Metering Project (SMP) dataset show that the proposed sensor-based detection framework achieves superior performance, with an accuracy of 88.45% and a false alarm rate of just 13.85%, while reducing training time by over 75%. By efficiently processing high-frequency smart meter sensor data, this model contributes to developing real-time and energy-efficient anomaly detection systems in smart grid environments.

Time-dependent scale-free brain dynamics during naturalistic inputs.

Environmental processes, such as auditory and visual inputs, often follow power-law distributions with a time-dependent and constantly changing spectral exponent, β(t). However, it remains unclear how the brain's scale-free dynamics continuously respond to naturalistic inputs, such as by potentially alternating instead of static levels of the spectral exponent. Our fMRI study investigates the brain's dynamic, time-dependent spectral exponent, β(t), during movie-watching, and uses time-varying inter-subject correlation, ISC(t), to assess the extent to which input dynamics are reflected as shared brain activity across subjects in early sensory regions. Notably, we investigate the level of ISC particularly based on the modulation by time-dependent scale-free dynamics or β(t). We obtained three key findings: First, the brain's β(t) showed a distinct temporal structure in visual and auditory regions during naturalistic inputs compared to the resting-state, investigated in the 7 Tesla Human Connectome Project dataset. Second, β(t) and ISC(t) were positively correlated during naturalistic inputs. Third, grouping subjects based on the Rest-to-Movie standard deviation change of the time-dependent spectral exponent β(t) revealed that the brain's relative shift from intrinsic to stimulus-driven scale-free dynamics modulates the level of shared brain activity, or ISC(t), and thus the imprinting of inputs on brain activity. This modulation was further supported by the observation that the two groups displayed significantly different β(t)-ISC(t) correlations, where the group with a higher mean of ISC(t) during inputs also exhibited a higher β(t)-ISC(t) correlation in visual and auditory regions. In summary, our fMRI study underscores a positive relationship between time-dependent scale-free dynamics and ISC, where higher spectral exponents correspond to higher degrees of shared brain activity during ongoing audiovisual inputs.

The Analysis on Local Network Roles: Evidence from Construction Contractors in China’s PPP Industry

Network role analysis has been a hot topic in the field of organizational networks because organizations that occupy a dominant network role tend to be more profitable and competitive. This study aims to analyze the local network roles based on structural equivalence in public-private partnership (PPP) project networks and explore how to occupy the local roles based on a resource-based view. This study employed structural equation modeling and tested the proposed hypotheses using a longitudinal data set of PPP transportation projects in China. The findings show that there are 11 types of local network roles identified in the project network, which can be categorized as star-center roles (four types), bridge roles (four types), and peripheral roles (three types). We observed that construction contractors (CCs) exhibit the highest intensity in star roles. Moreover, the financial capabilities of CCs significantly influence their occupation of star-center and bridge roles. Technical capabilities only support their assumption of the partial bridge and peripheral roles, while project management capabilities have an impact on assuming all types of bridge roles and part star-center roles. This study provides new insights into network role analysis, deepening the understanding of the micro-level behaviors of the project network.

Multi-level relationship patterns among private-sector organizations in public–private partnerships: a network perspective with evidence from China’s transportation sectors

PurposeThe nature of public-private partnership (PPP) markets, industry competition, and organizational embeddedness (e.g. collaboration experience), significantly influence how private-sector organizations in PPP establish relational ties and create project networks during the PPP procurement phase. This necessitates a multi-level perspective to unveil the relationship mechanisms and behaviors of organizations in PPP at different levels. This study aims to analyze the industry, clustering, and local characteristics of organization relationships using a longitudinal dataset of PPP transportation projects sourced from the National PPP Integrated Information Platform (NPPPIIP).Design/methodology/approachThis study adopts a multi-level analytical approach, encompassing whole-network, clustering-network, and local-network levels, to explore the mechanisms of organizational relationships and behavioral patterns within collaboration networks among private-sector organizations.FindingsThe results indicate that (1) organizational relationships within PPP unfold spontaneously over time, driven intrinsically and collectively, devoid of prior planning, control, external direction, or imposed order at the industry level, with the evolution of four collaboration types elucidating the emergence of self-organizing networks; (2) super enterprises exhibit exceptional breadth, brokerage, and ego abilities, and relationship patterns are governed by the effects of geographic proximity and preferential attachment at the clustering level; and (3) local relationship patterns are characterized by embeddedness mechanisms, including relational and structural aspects.Originality/valueThe multi-level analysis framework enhances our understanding of organization interactions in the PPP procurement phase, breaking away from the traditional social network analysis (SNA) at both the network-centric and ego-centric levels. Overall, the insights are valuable for policymakers designing improved tendering processes and for PPP private-sector organizations seeking a comprehensive overview of competition and collaboration in the transportation infrastructure sector.

Enhancing analogy-based software cost estimation using Grey Wolf Optimization algorithm.

Accurate software cost estimation (SCE) is a critical factor in the successful delivery of software projects, as highlighted by industry statistics indicating that only some of the projects comply with the predicted budget. Among the software estimation methods, analogy-based estimation (ABE) is one of the most popular ones. Although this method has been customized in recent years with the help of optimization algorithms to achieve better results, the use of more powerful optimization algorithms can be effective in achieving better results in software size estimation. This study presents an innovative approach to SCE that integrates the grey wolf optimization (GWO) algorithm to enhance the precision of ABE. The GWO algorithm, inspired by the hunting behavior and social hierarchy of grey wolves, is mathematically modeled and incorporated into the ABE approach. The key focus of this research is the optimization of the similarity function, a crucial component of the ABE, using both Euclidean and Manhattan distance measures. The article addresses the challenges in selecting an optimal similarity function and emphasizes the importance of proper feature weighting to improve estimation accuracy. The proposed GWO-based ABE method is rigorously evaluated on multiple software project datasets using cross-validation techniques. The performance of the GWO-based ABE is compared to other evolutionary algorithms based on widely accepted evaluation metrics. The results confirm that the integration of the GWO algorithm into ABE enhances estimation accuracy and model robustness. By optimizing feature weights in the similarity function, GWO-ABE effectively addresses key limitations of traditional analogy-based methods. The proposed approach demonstrates superior performance across multiple datasets, particularly under the Euclidean distance function, making it a reliable solution for software project cost estimation. Experimental evaluations show that GWO-ABE achieves notable improvements in key performance metrics, leading to reduced mean magnitude of relative error (MMRE), median magnitude of relative error (MdMRE), and higher percentage of prediction (PRED) compared to other ABE-customized methods. These findings highlight the role of metaheuristic optimization in improving software estimation techniques, contributing to more precise and efficient project planning and management.

Independent validation and outlier analysis of EuroPOND alzheimer’s disease staging model using ADNI and real-world clinical data

BackgroundEvent-based modeling (EBM) traces sequential progression of events in complex processes like neurodegenerative diseases, adept at handling uncertainties. This study validated an EBM for Alzheimer’s disease (AD) staging designed by EuroPOND, an EU-funded Horizon 2020 project, using research and real-world datasets, a crucial step towards application in multi-center trials.MethodsThe training dataset comprised 1737 subjects from ADNI-1/GO/2, using the EuroPOND EBM toolbox. Testing datasets included a research cohort from University of Antwerp (controls, CN (n = 46), subjective cognitive decline, SCD (n = 10), mild cognitive impairment, MCI (n = 47), AD dementia, ADD (n = 16)) and a real-world cohort from 9 Belgian Dementia Council memory clinics (CN (n = 91), SCD (n = 66), (non-amnestic) naMCI (n = 54), aMCI (n = 255), and ADD (n = 220). Biomarkers included: 2 clinical scores (Mini Mental State Examination (MMSE), Rey Auditory Verbal Learning Test (RAVLT)); 3 CSF-biomarkers (Aβ1−42, P-tau181, total-Tau); and 4 magnetic resonance imaging (MRI) biomarkers (volumes of the hippocampi, temporal, parietal, and frontal cortices) computed with icobrain dm. The naMCI and aMCI groups were compared by EBM stage proportions, and the model’s effectiveness at patient level was evaluated.ResultsThe research cohort’s maximum likelihood event sequence comprised CSF Aβ1-42, P-tau181, T-tau, RAVLT, MMSE, and cortical volumes. The clinical cohort’s order was frontal cortex volume, MMSE, and remaining cortical regions. aMCI subjects showed higher staging than naMCI, with 54% in the two most advanced stages compared to 38% in naMCI. In the research cohort, 10 outliers were identified with potential mismatches between assigned stages and clinical or biomarker profiles, with CN (n = 4) and SCD (n = 2) subjects assigned in stage 4, one control in stage 9 with abnormal imaging, and three aMCI cases in stage 0 despite clinical or volumetric signs of impairment.ConclusionsThis study highlights the generalizability of EuroPOND’s AD EBM model across research and real-world clinical datasets, supporting its use in multi-center trials. aMCI subjects generally reside in more advanced stages than naMCI, who may not necessarily have AD, demonstrating utility for precision recruitment/screening.

Computational Material Science Has a Data Problem.

We present an overdue questioning of the computational material science data: Is it suitable for training machine learning models? By examining the energy above the convex hull (Eh), the electronic bandgap, and the formation energy data in the Materials Project dataset, we find that Eh is an unsteady quantity, as are DFT-computed voltages, because the present materials in the database do not constitute sufficient representation of the chemical spaces that are necessary to account for crystal decomposition. We also show that there are discrepancies in the reported electronic bandgap values in the Materials Project database and that the formation energy data can potentially shift due to changing the optimization parameters that reduce the energy of the structure below the value deposited in the database.

Consumption quota compilation based on BP artificial neural network algorithm in mechanical and electrical installation engineering of prefabricated buildings

The traditional quota compilation method has a large workload and requires a lot of manpower and material resources, making it difficult to apply to the consumption quota compilation in mechanical and electrical installation engineering of prefabricated buildings. Therefore, a consumption quota compilation model on the basis of artificial neural network is built. On the basis of the traditional quota formulation model based on statistical theory, artificial neural networks are introduced, and regularization techniques and particle swarm optimization algorithms are taken to optimize the model performance. The experiment was validated using project datasets covering different regions, scales, and types of prefabricated components. The results showed that the mean squared errors on the training and testing sets were 1.2% and 1.1%, and the average absolute errors were 8.3% and 8.1%, respectively. In addition, the determination coefficients on the training and testing sets were 95.1% and 92.8%, and the accuracy was 92.3% and 91.4%. Further case analysis also showed that the prediction error rates of the research model for material consumption, labor hours, and mechanical equipment usage were relatively low, not exceeding 2.48%, 1.25%, and 4.1%, respectively. In addition, in terms of quota compilation efficiency and economic benefits, the proposed model achieved a quota compilation efficiency value of 90.1%. The return on investment in material consumption, labor hours, and mechanical equipment use was 5.03, 6.09, and 5.92, respectively, and the cost savings rates were 6.21%, 4.85%, and 5.48%, respectively, all of which were better than traditional models. Overall, the designed model can optimize the accuracy of engineering budgeting and the ability to control costs.

Integrative genome-wide association studies, proteome-wide Mendelian randomization, and single-cell RNA sequencing identify interleukin-6 receptor protein as a therapeutic target in aortic aneurysm.

Integrative genome-wide association studies, proteome-wide Mendelian randomization, and single-cell RNA sequencing identify interleukin-6 receptor protein as a therapeutic target in aortic aneurysm.

Improving delay and strength maps derived from resting-state fMRI using PCA-based denoising and group data from the HCP dataset.

Improving delay and strength maps derived from resting-state fMRI using PCA-based denoising and group data from the HCP dataset.

Dual-Uncertainty Guided Multimodal MRI-Based Visual Pathway Extraction.

This study aims to accurately extract the visual pathway (VP) from multimodal MR images while minimizing reliance on extensive labeled data and enhancing extraction performance. We propose a novel approach that incorporates a Modality-Relevant Feature Extraction Module (MRFEM) to effectively extract essential features from T1-weighted and fractional anisotropy (FA) images. Additionally, we implement a mean-teacher model integrated with dual uncertainty-aware ambiguity identification (DUAI) to enhance the reliability of the VP extraction process. Experiments conducted on the Human Connectome Project (HCP) and Multi-Shell Diffusion MRI (MDM) datasets demonstrate that our method reduces annotation efforts by at least one-third compared to fully supervised techniques while achieving superior extraction performance over six state-of-the-art semi-supervised methods. The proposed label-efficient approach alleviates the burdens of manual annotation and enhances the accuracy of multimodal MRI-based VP extraction. This work contributes to the field of medical imaging by facilitating more efficient and accurate visual pathway extraction, thereby improving the analysis and understanding of complex brain structures with reduced reliance on expert annotation.

Untamed: Unconstrained Tensor Decomposition and Graph Node Embedding for Cortical Parcellation.

Cortical parcellation is fundamental to neuroscience, enabling the division of cerebral cortex into distinct, non-overlapping regions to support interpretation and comparison of complex neuroimaging data. Although extensive literature has investigated cortical parcellation and its connection to functional brain networks, the optimal spatial features for deriving parcellations from resting-state fMRI (rsfMRI) remain unclear. Traditional methods such as Independent Component Analysis (ICA) have been widely used to identify large-scale functional networks, while other approaches define disjoint cortical parcellations. However, bridging these perspectives through effective feature extraction remains an open challenge. To address this, we introduce Untamed , a novel framework that integrates unconstrained tensor decomposition using NASCAR to identify functional networks, with state-of-the-art graph node embedding to generate cortical parcellations. Our method produces near-homogeneous, spatially coherent regions aligned with large-scale functional networks, while avoiding strong assumptions like statistical independence required in ICA. Across multiple datasets, Untamed consistently demonstrates improved or comparable performance in functional connectivity homogeneity and task contrast alignment compared to existing atlases. The pipeline is fully automated, allowing for rapid adaptation to new datasets and the generation of custom parcellations. The atlases derived from the Genomics Superstruct Project (GSP) dataset, along with the code for generating customizable parcel numbers, are publicly available at https://untamed-atlas.github.io .

Machine Learning Matches Human Performance at Segmenting the Human Visual Cortex

A major problem in human visual neuroscience research is the localization of visual areas on the cortical surface. Currently available methods are capable of making detailed predictions about many areas, but human raters do not agree as well with these methods as they do with each other. Although highly accurate, human raters require substantial time and expertise that many researchers do not have. Additionally, human raters require functional data for drawing visual area boundaries that requires additional scan time, budget, and expertise to collect. Here, we train convolutional neural network (CNN) models to predict the boundaries and iso-eccentric regions of V1, V2, and V3 in both the Human Connectome Project dataset and the NYU Retinotopy dataset. CNNs trained to use the same functional data available to human raters predicted these boundaries with an accuracy similar to humans, while CNNs trained to use only anatomical data had a lower accuracy that was nonetheless higher than that of any currently available method. A comparison of the model accuracies when predicting eccentricity-based boundaries and polar angle-based boundaries suggests that eccentricity is substantially less closely tied to anatomical structure than polar angle and that the cortical magnification function, at least in terms of eccentricity, varies substantially between subjects. We further find that the fraction of V1, V2, and V3 that can be accurately parcellated into function regions using gray matter structural data alone is ~75% (~80% of the inter-rater reliability of human experts), implying a much tighter coupling between structure and function in these areas than previously estimated. We conclude that machine learning techniques such as CNNs provide a powerful tool for mapping the brain with human accuracy and predict that such tools will become integral to neuroscience research going forward.

ScAI-SNP: a method for inferring ancestry from single-cell data

BackgroundCollaborative efforts, such as the Human Cell Atlas, are rapidly accumulating large amounts of single-cell data. To ensure that single-cell atlases are representative of human genetic diversity, we need to determine the ancestry of the donors from whom single-cell data are generated. Self-reporting of race and ethnicity, although important, can be biased and is not always available for the datasets already collected.MethodsHere, we introduce scAI-SNP, a tool to infer ancestry directly from single-cell genomics data. To train scAI-SNP, we identified 4.5 million ancestry-informative single-nucleotide polymorphisms (SNPs) in the 1000 Genomes Project dataset across 3201 individuals from 26 population groups. For a query single-cell dataset, scAI-SNP uses these ancestry-informative SNPs to compute the contribution of each of the 26 population groups to the ancestry of the donor from whom the cells were obtained.ResultsUsing diverse single-cell datasets with matched whole-genome sequencing data, we show that scAI-SNP is robust to the sparsity of single-cell data, can accurately and consistently infer ancestry from samples derived from diverse types of tissues and cancer cells, and can be applied to different modalities of single-cell profiling assays, such as single-cell RNA-seq and single-cell ATAC-seq.DiscussionFinally, we argue that ensuring that single-cell atlases represent diverse ancestry, ideally alongside race and ethnicity, is ultimately important for improved and equitable health outcomes by accounting for human diversity.

Terahertz transmission tomography enabled by focal plane array detection

This paper presents a terahertz transmission tomography system that integrates an optically pumped 2.52 THz laser with a high-speed focal plane array (FPA) detector for rapid three-dimensional imaging of small-scale samples. The FPA detector acquires a complete 180° projection dataset in just 8 min, significantly reducing imaging time compared to conventional point-by-point scanning systems. Projection data are reconstructed via a filtered back-projection algorithm optimized with a Hamming filter, effectively balancing noise suppression and structural fidelity. Quantitative evaluation shows that the Hamming filter achieves an SNR of 1.42 (versus 1.35–1.38 for the Ram–Lak and Shepp–Logan filters), maintains structural entropy at 6.36 (comparable to 6.36–6.37), and provides a 64% reduction in edge artifacts (Laplacian variance 0.0048 versus 0.0117). Three-dimensional reconstruction of a PTFE hose confirms the system’s ability to resolve multilayer structures with minimal distortion, underscoring its potential for non-destructive testing and biomedical imaging.

The subcortical correlates of self-reported sleep quality.

To assess the association between self-reported measures of sleep quality and cortical and subcortical local morphometry. Sleep quality, operationalized with the Pittsburgh Sleep Quality Index (PSQI), and neuroanatomical data from the full release of the young adult Human Connectome Project dataset were analyzed (N=1,112; 46% female; mean age: 28.8 years old). Local cortical and subcortical morphometry was measured with subject-specific segmentations resulting in voxelwise gray matter difference (i.e., voxel-based morphometry) measurements for cortex and local shape measurements for subcortical regions. Associations between the total score of PSQI, two statistical groupings of its subcomponents (obtained with a principal component analysis), and their interaction with demographic (i.e., sex, age, handedness, years of education) and biometric (i.e., BMI) variables were assessed using a general linear model and a nonparametric permutation approach. Sleep quality-related variance was significantly associated with subcortical morphometry, particularly in the bilateral caudate, putamen, and left pallidum, where smaller shape measures correlated with worse sleep quality. Notably, these associations were independent of demographic and biometric factors. In contrast, cortical morphometry, along with additional subcortical sites, showed no direct associations with sleep quality but demonstrated interactions with demographic and biometric variables. This study reveals a specific link between self-reported sleep quality and subcortical morphometry, particularly within the striatum and pallidum, reinforcing the role of these regions in sleep regulation. These findings underscore the importance of considering subcortical morphology in sleep research and highlight potential neuromodulatory targets for sleep-related interventions.

Asymmetric Convolution-based GAN Framework for Low-Dose CT Image Denoising.

Asymmetric Convolution-based GAN Framework for Low-Dose CT Image Denoising.

Predicting the number of bidders in construction competitive bidding using explainable machine learning models

PurposeThe number of bidders in upcoming tenders has important managerial implications for both construction clients and contractors in their decision-making in the competitive bidding process. However, there is a stagnation of research efforts on predicting the number of bidders with only a handful of studies over the past decades, which mainly focused on statistical distribution of the number of bidders. This study aims to provide a new perspective of predicting the number of bidders using machine learning (ML) algorithms.Design/methodology/approachThis study adopted a case study approach with a bidding dataset of public sector construction projects in Singapore. Six ML models were developed, and linear regression was used as a baseline model is assessing the predictive performance of ML models.FindingsThe results show that ML models outperform the baseline linear regression model, in which XGBoost is the best performing model of R2 which is two times higher than the linear regression model. In addition, economic-related factors play a vital role in this prediction problem.Research limitations/implicationsWhile the predictive performance of the developed ML models is relatively low, it indicates the challenges and complexities in this prediction problem, even with the use of artificial intelligent techniques.Originality/valueBeing a pioneering work, this study sets a foundation for the use of ML models in this prediction problem and offers insights for future modelling attempts towards the development of a decision support system for construction clients and contractors.

An extragradient and noise-tuning adaptive iterative network for diffusion MRI-based microstructural estimation.

An extragradient and noise-tuning adaptive iterative network for diffusion MRI-based microstructural estimation.

Class Imbalance-oriented Online Feature Selection Method for Just-in-time Software Defect Prediction

Just-in-time software defect prediction (JIT-SDP) is a defect prediction technique that targets changes in software code, offering significant advantages in quickly identifying potential defects and improving development efficiency. However, most existing methods assume that the importance of features remains stable over time, overlooking the dynamic changes in feature distributions and the evolution of class imbalance in real-world development environments. This limitation eventually degrades the predictive performance. To address this issue, this paper proposes an Imbalance-oriented Online Feature Selection (IOFS) method, which dynamically adjusts the feature importance and uncertainty parameters to adapt in real time to concept drift and class imbalance in data streams, thereby enhancing model performance and generalization. The experimental validation on 14 open-source project datasets demonstrates that IOFS significantly improves the values of [Formula: see text]Mean on 11 datasets and effectively reduces the average of the absolute differences between recalls for each time step, exhibiting robustness to dynamic feature changes and sensitivity to development-phase feature differences. This study provides an effective solution for online JIT-SDP.