Connection Topology Research Articles

A comprehensive hybrid transient CFD-thermal resistance model for automobile thermoelectric generators

This paper proposes a comprehensive hybrid transient CFD-thermal resistance model to predict the dynamic behaviour of an automobile thermoelectric generator (ATEG) system. The model takes into account the temperature dependences, the topological connection of thermoelectric modules, and the dynamic characteristics, which has the merits of high accuracy and short computational time. The dynamic behaviour of the ATEG system is determined and thoroughly examined using the transient exhaust heat as the heat source input. According to the transient model results, the dynamic output power of the ATEG system keeps the same variation trend with the exhaust temperature, but the variation of output power is more stable. Under the whole driving cycle, the mean power and efficiency of the 1/4 ATEG system are 8.91 W and 3.39% respectively, which are 3.39% lower and 47.52% higher than those expected by steady-state analysis. Beside, the model is validated experimentally, and the mean deviations of the output voltage and outlet air temperature are 7.70% and 1.12% respectively. This model is convenient to evaluate the behaviour of the ATEG system under different topological connections and gives a fresh tool for assessing the dynamic behaviour of ATEG systems.

Research Progress on Interface Circuit of Capacitive Micro Accelerometer

Micro-Electro-Mechanical System (MEMS) capacitive accelerometers have received extensive attention in recent years due to their excellent performance indicators; especially in the fields of noise, power consumption, and bias instability, great development and progress have been made. In the field of noise, effective noise reduction is achieved by introducing oversampling modulation technology combined with digital noise reduction technology. In power consumption, the power consumption of the accelerometer is effectively reduced by using the successive approximation structure in the interface circuit and using the finite state machine for precise control. In bias instability, the effects of temperature offset and zero-point drift are suppressed by using a hybrid topology connection structure in the interface circuit, and an effective reduction of bias instability is achieved. In this paper, the research and progress of MEMS capacitive accelerometer in the field of noise, power consumption and bias instability are reviewed, and the articles published in recent years are listed and summarized and prospected.

On Some Meta-Theoretic Topological Features of the Region Connection Calculus

This paper examines several intended topological features of the Region Connection Calculus (RCC) and argues that they are either underdetermined by the formal theory or given by the complement axiom. Conditions are identified under which the axioms of RCC are satisfied in topological models under various set restrictions. The results generalise previous results in the literature to non-strict topological models and across possible interpretations of connection. It is shown that the intended interpretation of connection and the alignment of self-connection with topological connection are underdetermined by the axioms of RCC, which suggests that additional axioms are necessary to secure these features. It is also argued that the complement axiom gives RCC models much of their topological structure. In particular, the incompatibility of RCC with interiors is argued to be given by the complement axiom.

Towards Mitigating Jellyfish Attacks Based on Honesty Metrics in V2X Autonomous Networks

In vehicle-to-everything (V2X) networks, security and safety are inherently difficult tasks due to the distinct characteristics of such networks, such as their highly dynamic topology and frequent connectivity disruptions. Jellyfish attacks are a sort of denial of service attack that are challenging to deal with, since they conform to protocol norms while impairing network performance, particularly in terms of communication overhead and reliability. Numerous existing approaches have developed new techniques with which to identify and prevent these attacks; however, no approach has been capable of facing all three types of Jellyfish attacks, which include reordering attacks, delay variance attacks, and periodic drop attacks. In this work, we design a new protocol that analyzes the behavior of every node in a network and selects the trusted routes for data transmission to their intended destination by calculating different Honesty metrics. The OMNET++ simulator was used to evaluate the overall performance of the proposed protocol. Various evaluation metrics, such as the packet delivery ratio, end-to-end delay, and throughput, are considered and compared with other existing approaches.

Excess BMI in early adolescence adversely impacts maturating functional circuits supporting high-level cognition and their structural correlates.

Adverse effects of excess BMI (affecting 1 in 5 children in the US) on brain circuits during neurodevelopmentally vulnerable periods are incompletely understood. This study investigated BMI-related alterations in maturating functional networks and their underlying brain structures, and high-level cognition in early adolescence. Cross-sectional resting-state fMRI, structural sMRI, neurocognitive task scores, and BMI from 4922 youth [median (IQR) age = 120.0 (13.0) months, 2572 females (52.25%)] from the Adolescent Brain Cognitive Development (ABCD) cohort were analyzed. Comprehensive topological and morphometric network properties were estimated from fMRI and sMRI, respectively. Cross-validated linear regression models assessed correlations with BMI. Results were reproduced across multiple fMRI datasets. Almost 30% of youth had excess BMI, including 736 (15.0%) with overweight and 672 (13.7%) with obesity, and statistically more Black and Hispanic compared to white, Asian and non-Hispanic youth (p < 0.01). Those with obesity or overweight were less physically active, slept less than recommended, snored more frequently, and spent more time using an electronic device (p < 0.01). They also had lower topological efficiency, resilience, connectivity, connectedness and clustering in Default-Mode, dorsal attention, salience, control, limbic, and reward networks (p ≤ 0.04, Cohen's d: 0.07-0.39). Lower cortico-thalamic efficiency and connectivity were estimated only in youth with obesity (p < 0.01, Cohen's d: 0.09-0.19). Both groups had lower cortical thickness, volume and white matter intensity in these networks' constituent structures, particularly anterior cingulate, entorhinal, prefrontal, and lateral occipital cortices (p < 0.01, Cohen's d: 0.12-0.30), which also mediated inverse relationships between BMI and regional functional topologies. Youth with obesity or overweight had lower scores in a task measuring fluid reasoning - a core aspect of cognitive function, which were partially correlated with topological changes (p ≤ 0.04). Excess BMI in early adolescence may be associated with profound aberrant topological alterations in maturating functional circuits and underdeveloped brain structures that adversely impact core aspects of cognitive function.

Knowledge-Driven Cybersecurity Intelligence: Software Vulnerability Coexploitation Behavior Discovery

Coexploitation behavior, referring to multiple software vulnerabilities being exploited jointly by one or more exploits, brings enormous challenges to the prevention and remediation of cyberattacks. Leveraging the latest advances in graph-driven intelligence, this article formulates vulnerability coexploitation behavior discovery as a link prediction problem between vulnerability entities within a vulnerability knowledge graph. We propose a modality-aware graph convolutional network (MAGCN) module to embed multimodality entity attributes and topological graph connectivity features into a unified lower dimensional feature space to boost link prediction performance. We further design a graph knowledge transfer learning (GKTL) strategy to transfer knowledge between subgraphs extracted from the same knowledge graph. Experimental results on a real-world dataset containing coexploitation incidents between 1995 and 2021 show that MAGCN achieved 81.34% on the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">F</i> 1 score when applying the GKTL strategy, superior to other graph neural network modules, such as GCN, GraphSAGE, EdgeGCN, and GINGCN.

Predicting Drug-Target Affinity by Learning Protein Knowledge From Biological Networks.

Predicting drug-target affinity (DTA) is a crucial step in the process of drug discovery. Efficient and accurate prediction of DTA would greatly reduce the time and economic cost of new drug development, which has encouraged the emergence of a large number of deep learning-based DTA prediction methods. In terms of the representation of target proteins, current methods can be classified into 1D sequence- and 2D-protein graph-based methods. However, both two approaches focused only on the inherent properties of the target protein, but neglected the broad prior knowledge regarding protein interactions that have been clearly elucidated in past decades. Aiming at the above issue, this work presents an end-to-end DTA prediction method named MSF-DTA (Multi-Source Feature Fusion-based Drug-Target Affinity). The contributions can be summarized as follows. First, MSF-DTA adopts a novel "neighboring feature"-based protein representation. Instead of utilizing only the inherent features of a target protein, MSF-DTA gathers additional information for the target protein from its biologically related "neighboring" proteins in PPI (i.e., protein-protein interaction) and SSN (i.e., sequence similarity) networks to get prior knowledge. Second, the representation was learned using an advanced graph pre-training framework, VGAE, which could not only gather node features but also learn topological connections, therefore contributing to a richer protein representation and benefiting the downstream DTA prediction task. This study provides new perspective for the DTA prediction task, and evaluation results demonstrated that MSF-DTA obtained superior performances compared to current state-of-the-art methods.

Topological invariant and domain connectivity in moiré materials

Recently, a moir\'e material has been proposed in which multiple domains of different topological phases appear in the moir\'e unit cell due to a large moir\'e modulation. Topological properties of such moir\'e materials may differ from those of the original untwisted layered material. In this paper, we study how the topological properties are determined in moir\'e materials with multiple topological domains. We show a correspondence between the topological invariant of moir\'e materials at the Fermi level and the topology of the domain structure in real space. We also find a bulk-edge correspondence that is compatible with a continuous change in the truncation condition, which is specific to moir\'e materials. We demonstrate these correspondences in the twisted Bernevig-Hughes-Zhang model that describes general ${\mathbb{Z}}_{2}$-indicated topological insulator phases, by tuning its domain structure in the moir\'e unit cell. The obtained result can only be used to determine the topology at the Fermi level in the charge-neutral condition, but it can be combined with the traditional Wilson loop analysis to determine topologies of other gaps around. These results give a feasible method to evaluate a topological invariant for all occupied bands of a moir\'e material and contribute to the design of topological moir\'e materials and devices.

Defense against underwater spy-robots: A distributed anti-theft topology control mechanism for insecure UASN

Underwater Acoustic Sensor Network (UASN) is deployed for various underwater applications, such as underwater creature tracking and underwater tactical surveillance. Particularly, UASN in military applications could be invaded by some underwater spy-robots termed eavesdroppers. The eavesdroppers navigate around some anchored nodes of UASN and eavesdrop on their communication channels silently. The anchored nodes are difficult to perceive the adjacent eavesdroppers, because the eavesdroppers never actively communicate with others. Thus, the anchored nodes could disseminate the data messages while they are blithely unaware of the adjacent eavesdroppers capturing the data messages. To reduce the theft ratio of data messages and guarantee the topology connectivity of UASN, an analysis framework regarding the theft ratio of data messages is first formulated, and then a geometric distribution is constructed for the in-degrees of anchored nodes. Furthermore, a binomial distribution for the communication ranges of anchored nodes is derived from this analysis framework. In our proposed Anti-Theft Topology Control Mechanism (ATTCM), the anchored nodes set the initial communication ranges according to the obtained binomial distribution, and then the communication ranges of anchored nodes are checked and adjusted to guarantee the topology connectivity. Simulation results demonstrate the superior performance of ATTCM, i.e., ATTCM can reduce the theft ratio of data messages effectively, and it can guarantee that there is at least one available communication path from each anchored node to the surface sink.

A global prediction of the Kataura plot for chiral carbon nanotubes: Topological family effect revealed in the natural helical crystal lattice scheme

The optical transition energies of chiral carbon nanotubes (CNTs) are intimately related to their geometrical structures, and are frequently applied in the recognition and isolation of single nanotubes. Conventional translational crystal lattice (TCL) scheme failed to offer an accurate description of the optical properties of the chiral CNTs due to the formidable unit cell size constraint as well as the complexity of the band structures. Most fitting equations for the Kataura plot, usually based on the empirical correlations with the chiral indices of the CNTs, lack a clear physical interpretation. In this study, we apply the natural helical crystal lattice (NHCL) scheme to systematically analyze the band gap variation of arbitrary chiral CNTs. We derive a global analytical formula that takes into account the comprehensive geometrical connection topologies for the band gap of all (n, m)-CNTs with a diameter larger than 1.2 nm. With the inclusion of the many-body effect, we are able to provide a global Kataura plot prediction with a minimal set of only two fitting parameters, and the results are consistent with the experimental Kataura plot within 0.67% deviation for over 1300 data points. Our formula precisely captures the observed experimental branching features and presents a clear physical rationalization for the topological family effect on the optical excitation energies of chiral CNTs.

Exploring the Restoration of Brain Connectivity during Weight Normalization in Severe Anorexia Nervosa

Abstract Anorexia nervosa is a persistent and often difficult to treat eating disorder with significant physical and mental health consequences. While it is known that the disorder is associated with alterations in brain functional connectivity during the phase of acute underweight, the effect of weight normalization on brain connectivity remains unclear. This talk focuses on the recovery of intrinsic brain connectivity during weight normalization in severe anorexia nervosa, presenting data from a longitudinal study. Using resting-state functional magnetic resonance imaging, we assessed brain connectivity at three different stages of inpatient treatment. Our findings indicate that patients with severe anorexia nervosa have weaker intrinsic connectivity and altered network topology, which do not improve during treatment. These persistent disruptions in brain networks suggest that severe anorexia nervosa may have long-term effects on the way the brain processes information, even after weight is restored. Disclosure of Interest None Declared

Molecular Tracking of the Origin of Vesicular Stomatitis Outbreaks in 2004 and 2018, Ecuador

The Vesicular Stomatitis Virus (VSV) is an arbovirus causing vesicular stomatitis (VS) in livestock. There are two serotypes recognized: New Jersey (VSNJV) and Indiana (VSIV). The virus can be transmitted directly by contact or by vectors. In 2018, Ecuador experienced an outbreak of Vesicular Stomatitis (VS) in cattle, caused by VSNJV and VSVIV, with 399 cases reported distributed over 18 provinces. We determined the phylogenetic relationships among 67 strains. For the construction of phylogenetic trees, the viral phosphoprotein gene was sequenced, and trees were constructed based on the Maximum Likelihood method using 2004 outbreak strains from Ecuador (GenBank) and the 2018 sequences (this article). We built a haplotype network for VSNJV to trace the origin of the 2004 and 2018 epizootics through topology and mutation connections. These analyses suggest two different origins, one related to the 2004 outbreak and the other from a transmission source in 2018. Our analysis also suggests different transmission patterns; several small and independent outbreaks, most probably transmitted by vectors in the Amazon, and another outbreak caused by the movement of livestock in the Andean and Coastal regions. We recommend further research into vectors and vertebrate reservoirs in Ecuador to clarify the mechanisms of the reemergence of the virus.

MultiSimNeNc: A network representation learning-based module identification method by network embedding and clustering

Accurate identification of gene modules based on biological networks is an effective approach to understanding gene patterns of cancer from a module-level perspective. However, most graph clustering algorithms just consider low-order topological connectivity, which limits their accuracy in gene module identification. In this study, we propose a novel network-based method, MultiSimNeNc, to identify modules in various types of networks by integrating network representation learning (NRL) and clustering algorithms. In this method, we first obtain the multi-order similarity of the network using graph convolution (GC). Then, we aggregate the multi-order similarity to characterize the network structure and use non-negative matrix factorization (NMF) to achieve low-dimensional node characterization. Finally, we predict the number of modules based on the bayesian information criterion (BIC) and use the gaussian mixture model (GMM) to identify modules. To testify to the efficacy of MultiSimeNc in module identification, we apply this method to two types of biological networks and six benchmark networks, where the biological networks are constructed based on the fusion of multi-omics data from glioblastoma (GBM). The analysis shows that MultiSimNeNc outperforms several state-of-the-art module identification algorithms in identification accuracy, which is an effective method for understanding biomolecular mechanisms of pathogenesis from a module-level perspective.

Functional brain network alterations in the co-occurrence of autism spectrum disorder and attention deficit hyperactivity disorder.

Autism spectrum disorder (ASD) and attention deficit hyperactivity disorder (ADHD) are two highly prevalent and commonly co-occurring neurodevelopmental disorders. The neural mechanisms underpinning the comorbidity of ASD and ADHD (ASD + ADHD) remain unclear. We focused on the topological organization and functional connectivity of brain networks in ASD + ADHD patients versus ASD patients without ADHD (ASD-only). Resting-state functional magnetic resonance imaging (rs-fMRI) data from 114 ASD and 161 typically developing (TD) individuals were obtained from the Autism Brain Imaging Data Exchange II. The ASD patients comprised 40 ASD + ADHD and 74 ASD-only individuals. We constructed functional brain networks for each group and performed graph-theory and network-based statistic (NBS) analyses. Group differences between ASD + ADHD and ASD-only were analyzed at three levels: nodal, global, and connectivity. At the nodal level, ASD + ADHD exhibited topological disorganization in the temporal and occipital regions, compared with ASD-only. At the global level, ASD + ADHD and ASD-only displayed no significant differences. At the connectivity level, the NBS analysis revealed that ASD + ADHD showed enhanced functional connectivity between the prefrontal and frontoparietal regions, as well as between the orbitofrontal and occipital regions, compared with ASD-only. The hippocampus was the shared region in aberrant functional connectivity patterns in ASD + ADHD and ASD-only compared with TD. These findings suggests that ASD + ADHD displays altered topology and functional connectivity in the brain regions that undertake social cognition, language processing, and sensory processing.

Automatic vessel crossing and bifurcation detection based on multi-attention network vessel segmentation and directed graph search

Analysis of the vascular tree is the basic premise to automatically diagnose retinal biomarkers associated with ophthalmic and systemic diseases, among which accurate identification of intersection and bifurcation points is quite challenging but important for disentangling complex vascular network and tracking vessel morphology. In this paper, we present a novel directed graph search-based multi-attentive neural network approach to automatically segment the vascular network and separate intersections and bifurcations from color fundus images. Our approach uses multi-dimensional attention to adaptively integrate local features and their global dependencies while learning to focus on target structures at different scales to generate binary vascular maps. A directed graphical representation of the vascular network is constructed to represent the topology and spatial connectivity of the vascular structures. Using local geometric information including color difference, diameter, and angle, the complex vascular tree is decomposed into multiple sub-trees to finally classify and label vascular feature points. The proposed method has been tested on the DRIVE dataset and the IOSTAR dataset containing 40 images and 30 images, respectively, with 0.863 and 0.764 F1-score of detection points and average accuracy of 0.914 and 0.854 for classification points. These results demonstrate the superiority of our proposed method outperforming state-of-the-art methods in feature point detection and classification.

Exploring the neuroprotective effects of montelukast on brain inflammation and metabolism in a rat model of quinolinic acid-induced striatal neurotoxicity

BackgroundOne intrastriatal administration of quinolinic acid (QA) in rats induces a lesion with features resembling those observed in Huntington’s disease. Our aim is to evaluate the effects of the cysteinyl leukotriene receptor antagonist montelukast (MLK), which exhibited neuroprotection in different preclinical models of neurodegeneration, on QA-induced neuroinflammation and regional metabolic functions.MethodsThe right and left striatum of Sprague Dawley and athymic nude rats were injected with QA and vehicle (VEH), respectively. Starting from the day before QA injection, animals were treated with 1 or 10 mg/kg of MLK or VEH for 14 days. At 14 and 30 days post-lesion, animals were monitored with magnetic resonance imaging (MRI) and positron emission tomography (PET) using [18F]-VC701, a translocator protein (TSPO)-specific radiotracer. Striatal neuroinflammatory response was measured post-mortem in rats treated with 1 mg/kg of MLK by immunofluorescence. Rats treated with 10 mg/kg of MLK also underwent a [18F]-FDG PET study at baseline and 4 months after lesion. [18F]-FDG PET data were then used to assess metabolic connectivity between brain regions by applying a covariance analysis method.ResultsMLK treatment was not able to reduce the QA-induced increase in striatal TSPO PET signal and MRI lesion volume, where we only detected a trend towards reduction in animals treated with 10 mg/kg of MLK. Post-mortem immunofluorescence analysis revealed that MLK attenuated the increase in striatal markers of astrogliosis and activated microglia in the lesioned hemisphere. We also found a significant increase in a marker of anti-inflammatory activity (MannR) and a trend towards reduction in a marker of pro-inflammatory activity (iNOS) in the lesioned striatum of MLK—compared to VEH-treated rats. [18F]-FDG uptake was significantly reduced in the striatum and ipsilesional cortical regions of VEH-treated rats at 4 months after lesion. MLK administration preserved glucose metabolism in these cortical regions, but not in the striatum. Finally, MLK was able to counteract changes in metabolic connectivity and measures of network topology induced by QA, in both lesioned and non-lesioned hemispheres.ConclusionsOverall, MLK treatment produced a significant neuroprotective effect by reducing neuroinflammation assessed by immunofluorescence and preserving regional brain metabolism and metabolic connectivity from QA-induced neurotoxicity in cortical and subcortical regions.

Learning multi-view and centerline topology connectivity information for pulmonary artery–vein separation

Background:Automatic pulmonary artery–vein separation has considerable importance in the diagnosis and treatment of lung diseases. However, insufficient connectivity and spatial inconsistency have always been the problems of artery–vein separation. Methods:A novel automatic method for artery–vein separation in CT images is presented in this work. Specifically, a multi-scale information aggregated network (MSIA-Net) including multi-scale fusion blocks and deep supervision, is proposed to learn the features of artery–vein and aggregate additional semantic information, respectively. The proposed method integrates nine MSIA-Net models for artery–vein separation, vessel segmentation, and centerline separation tasks along with axial, coronal, and sagittal multi-view slices. First, the preliminary artery–vein separation results are obtained by the proposed multi-view fusion strategy (MVFS). Then, centerline correction algorithm (CCA) is used to correct the preliminary results of artery–vein separation by the centerline separation results. Finally, the vessel segmentation results are utilized to reconstruct the artery–vein morphology. In addition, weighted cross-entropy and dice loss are employed to solve the class imbalance problem. Results:We constructed 50 manually labeled contrast-enhanced computed CT scans for five-fold cross-validation, and experimental results demonstrated that our method achieves superior segmentation performance of 97.7%, 85.1%, and 84.9% on ACC, Pre, and DSC, respectively. Additionally, a series of ablation studies demonstrate the effectiveness of the proposed components. Conclusion:The proposed method can effectively solve the problem of insufficient vascular connectivity and correct the spatial inconsistency of artery–vein.

Probabilistic Network Topology Prediction for Active Planning: An Adaptive Algorithm and Application

This article tackles the problem of active planning to achieve cooperative localization for multirobot systems under measurement uncertainty in GNSS-limited scenarios. Specifically, we address the issue of accurately predicting the probability of a future connection between two robots equipped with range-based measurement devices. Due to the limited range of the equipped sensors, edges in the network connection topology will be created or destroyed as the robots move with respect to one another. Accurately predicting the future existence of an edge, given imperfect state estimation and noisy actuation, is therefore a challenging task. An adaptive power series expansion (or APSE) algorithm is developed based on current estimates and control candidates. Such an algorithm applies the power series expansion formula of the quadratic positive form in a normal distribution. Finite-term approximation is made to realize the computational tractability. Further analyses are presented to show that the truncation error in the finite-term approximation can be theoretically reduced to a desired threshold by adaptively choosing the summation degree of the power series. Several sufficient conditions are rigorously derived as the selection principles. Finally, extensive simulation results and comparisons, with respect to both single and multirobot cases, validate that a formally computed and therefore more accurate probability of future topology can help improve the performance of active planning under uncertainty.

Impulsive consensus of stochastic multi-agent systems under semi-Markovian switching topologies and application

This paper studied the impulsive consensus of stochastic multi-agent systems (MASs) under semi-Markovian switching topologies. The semi-Markovian process is employed to describe the random changes of communication topologies among agents. Applying the impulsive control method, the communication among agents is only needed at impulsive control instants. For the connectivity of communication topologies, we only require that partial communication topologies contain a directed spanning tree, which remove the restriction that every communication topology contains a directed spanning tree in existing works. Using the Lyapunov function method and the stationary distribution of semi-Markovian process, the impulsive control protocol (ICP) is designed to ensure the almost surely exponential consensus of stochastic MASs. As an application, the obtained result is used to solve the secure impulsive consensus of stochastic MASs under random cyber-attacks. We proposed a novel model of random cyber-attacks, in which adversaries can destroy communication links of agents and durations of attack-active interval and attack-free interval are random variables. A simulation example is given to show the effectiveness of obtained results.

A Novel Graph-Based Framework for Classifying Urban Functional Zones with Multisource Data and Human Mobility Patterns

Recent research has shown the advantages of incorporating multisource geospatial data into the classification of urban functional zones (UFZs), particularly remote sensing and social sensing data. However, the effects of combining datasets of varying quality have not been thoroughly analyzed. In addition, human mobility patterns from social sensing data, which capture signals of human activities, are often represented by origin-destination pairs, thus ignoring spatial relationships between UFZs embedded in mobility trajectories. To address the aforementioned issues, this study proposed a graph-based UFZ classification framework that fuses semantic features from high spatial resolution (HSR) remote sensing images, points of interest, and GPS trajectory data. The framework involves three main steps: (1) High-level scene information in HSR remote sensing imageries was extracted through deep neural networks, and multisource semantic embeddings were constructed based on physical features and social sensing features from multiple geospatial data sources; (2) UFZ mobility graph was constructed by spatially joining trajectory information with UFZs to construct topological connections between functional parcel segments; and (3) UFZ segments and multisource semantic features were transformed into nodes and embeddings in the mobility graphs, and subsequently graph-based models were adopted to identify UFZs. The proposed framework was tested on Zhuhai and Singapore datasets. Results indicated that it outperformed traditional classification methods with an overall accuracy of 76.7% and 84.5% for Zhuhai and Singapore datasets, respectively. The proposed framework contributes to literature in heterogeneous data fusion and is generalizable to other UFZ classification scenarios where human mobility patterns play a role.