Inconsistency Issues Research Articles

Comparative assessment of SoilGrids system with regional soil data for advancing ecosystem reporting in Bulgaria

The aim of this study is to test the suitability of the SoilGrids system for ecosystem reporting, research and monitoring. The study is conducted in the Rila Mountains in Bulgaria, an area characterised by diverse ecological factors. We propose a methodological approach to compare SoilGrids predictions with independent point observations, addressing issues of inconsistency across survey layers when combining data from different sources. The comparative analysis is discussed in respect to point data, soil type, altitude and climate. The results show that the SoilGrids represents the main soil parameters well in terms of their dynamics over the altitudinal range. There is a good agreement between the observed and predicted values for the averages of the parameters - bulk density, coarse fraction, soil organic carbon (SOC) content and SOC stock. The average measured SOC stock (0-30 cm) is 58.54 t/ha, while the average predicted SOC stock (0-30 cm) is 55.38 t/ha. However, the study also showed that the predicted values for nitrogen content are almost two times higher than the observed figures and the pH values from the SoilGrids are less acidic than those measured in the field.

A multi-mode self-centering piezo actuator: design, analysis and its application in full-stroke microscopic imaging

Parasitic piezo dive is a primary method for cross-scale positioning, which can adjust the locking force appropriately to increase the service life but bring radial fluctuations, backward motion, and bi-directional inconsistency issues. In this work, a multi-mode self-centering piezo actuator is proposed, a theoretical analysis model of the transmission mechanism is established, and parameterized design is conducted using the finite element method. The actuator features normal pressure adjustment, parallel drive, parasitic drive, and rectangular drive modes. Under the first mode, a bias voltage of 50 Vp-p achieved a locking force adjustment of 2.5 N. Furthermore, the characteristics of the remaining modes were compared. It was observed that the parallel trajectory drive mode exhibited a higher velocity of 31.68 mm/s, while the rectangular trajectory drive mode suppressed the backward motion to achieve smooth driving, with a linear correlation coefficient R2 of the displacement reaching 0.9999. Compared to the traditional parasitic actuator, the radial fluctuations in these two drive modes were reduced by up to 98.4 %, and the bidirectional consistency were improved to 97 %. Finally, we completed convenient imaging of various specimens across the full-stroke of microscopes by integrated circuit design and Bluetooth communication on mobile phones.

Chinese utility weights for the EORTC cancer-specific utility instrument QLU-C10D.

The aim of this study is to provide Chinese utility weights for the European Organization for Research and Treatment of Cancer Quality of Life Utility Measure-Core 10 Dimensions (EORTC QLU-C10D) which is a preference-based cancer-specific utility instrument derived from the EORTC QLQ-C30. We conducted an online survey of the general population in China, with quota sampling for age and gender. Each respondent was asked to complete a discrete choice experimental survey consisting of 16 randomly selected choice sets. The conditional logit model and mixed logit model were used to analyze respondents' preferences, and the goodness of fit of the model was tested. A total of 2003 respondents were included in the analysis. Utility decrements within dimensions were typically monotonic. Monotonic inconsistency issues in the Fatigue, Sleep, and Nausea dimensions were normalized by monotonicity correction. Physical functioning, Pain, and Role functioning were associated with the greatest utility weights, with the smallest decrements being in Bowel problems and Emotional functioning. The utility value for the worst health state was 0.083, i.e. slightly higher than being dead. This study provides the first China-specific set of value for the QLU-C10D based on societal preferences of the Chinese adult general population. The value set can be used as a cancer-specific scoring system for economic evaluations of new oncology therapies and technologies in China.

Modified Epoxy Resin on the Burning Behavior and Mechanical Properties of Aramid Fiber Composite.

Aramid fiber/epoxy resin (AF/EP) composite has been heavily used as an impact protection material due to its excellent mechanical properties and lightweight merits. Meanwhile, it is also necessary to concern the flammability of matrix resin and the wick effect of aramid fiber, which would constitute a fire risk in harsh environments. In this work, a multifunctional flame-retardant modifier (EAD) was incorporated into the AF/EP system to improve the flame retardation. The addition of 5 wt% EAD made the AF/EP composite exhibit a high limiting oxygen index (LOI) value of 37.5%, self-extinguishment, as well as decreased total heat release and total smoke release. The results from thermogravimetric analysis (TGA) and dynamic mechanical analysis (DMA) demonstrated that the treated composites maintained good thermal stability. Due to the combined action of covalent and noncovalent bonds in the matrix-rich region, the interfacial bonding improved, which endowed AF/EP composite with strengthening and toughening effects. Compared with the control sample AF/EP, the tensile strength and ballistic parameter (V50) of the sample with 5 wt% EAD increased by 17% and 10%, accompanied with ductile failure mode. Furthermore, the flame-retardant mechanism was obtained by analyzing the actions in condensed and gaseous phases. Thanks to good compatibility and interfacial adhesion, the incorporation of EAD solved the inconsistent issue between flame retardancy and mechanical properties, which further expanded the application of AF/EP composite in the protection field.

Exploration of Language Landscape and Optimization Strategies in Beijing's South Luogu Lane

South Luogu Lane is the oldest neighborhood in Beijing, preserving the architectural style of the Yuan Dynasty's alleyways, attracting a large number of domestic and foreign tourists. This study focuses on the official and private language landscapes of its main street. Through field visits and photographic documentation, the authors found that the national language dominates, accompanied by English translations, facilitating foreign tourists. However, there are issues of inconsistency and lack of uniformity in the language landscape, such as inconsistent positioning and font sizes. These problems affect the cultural image of South Luogu Lane. Therefore, the authors propose optimization strategies to strengthen the protection of the cultural landscape and promote the sustainable development of South Luogu Lane. The authors hope that these measures will enhance the cultural value of South Luogu Lane, provide a better experience for tourists, and promote its influence as a cultural symbol of Beijing.

A refined and concise model of indices for quantitatively measuring lexical richness of Chinese university students’ EFL writing

In the existing literature, scholars have proposed various indices to measure the lexical richness (LR) of English as a foreign language (EFL) writing. However, there are currently issues of redundant indices and inconsistent usage. Attempting to address the research question of which indices are the most sensitive and effective ones to distinguish between different grade levels of Chinese university students’ EFL writing, this study aims to put forward a refined and concise model of indices that can truthfully reflect LR in EFL writing. A total of 180 compositions were selected from a Chinese EFL learner corpus: <i>Spoken and written English corpus of Chinese learners</i>. Scores of 28 LR indices of these compositions were computed using the software <i>Lexical Complexity Analyzer</i>,<i> MATTR</i>,<i> </i>and<i> Coh-Metrix</i>. One-way ANOVA or Welch’s ANOVA, depending on the variable’s homogeneity of variances, was conducted for each index. Two criteria were applied to determine which index of a measure should be included in the refined model: whether the difference of an index is significant among different grade levels and the effect size of ANOVA. Based on the quantitative results of ANOVAs and qualitative human judgment based on literature, six indices of the six LR measures were included in the refined model: lexical density, lexical sophistication-I, verb sophistication-II, number of different words-expected sequence 50, corrected TTR, and squared verb variation-I. This refined model addresses the issues of redundancy and inconsistency in previous studies, providing a more accurate and efficient tool for assessing LR in EFL writing.

An Approach for Identifying Dynamic Parameters in Robotic Systems with Inconsistent Joint Measurements

Abstract Achieving accurate robot control and realistic robot simulation relies on the precise modeling of robotic dynamics. Although the identification method for obtaining dynamic parameters has been developed for several years, joint inconsistency has largely been disregarded in prior studies. The inconsistency of joint actuators results in varying confidence levels of their measurement results, leading to the departure of the final identification parameters from valid values, thereby impacting the control performance. This paper presents a novel identification method that effectively addresses the issue of joint inconsistency by assigning distinct weights to each joint. The presented approach extends the least square (LS) method and incorporates the particle swarm optimization (PSO) algorithm to calculate the weights of each individual joint. This approach is referred to as PWLS (PSO-based weighted least square). Simulation experiments demonstrated the superior identification accuracy of the PWLS method compared to the LS method in a robot system characterized by joint inconsistency. Moreover, the experiments were performed on a 3-degree-of-freedom (DoFs) robotic limb, which exhibited improved identification performance in both the excitation and verification trajectories. These findings have promising implications for enhancing the control and simulation of robotic systems.

UniQRNet: Unifying Referring Expression Grounding and Segmentation with QRNet

Referring expression comprehension aims to align natural language queries with visual scenes, which requires establishing fine-grained correspondence between vision and language. This has important applications in multi-modal reasoning systems. Existing methods typically use text-agnostic visual backbones to extract features independently without considering the specific text input. However, we argue that the extracted visual features can be inconsistent with the referring expression, which hurts multi-modal understanding. To address this, we first propose Query-modulated Refinement Network (QRNet) that leverages language guidance to guide visual feature extraction. However, it only focuses on the grounding task that can only provide coarse-grained annotations in the form of bounding box coordinates. The guidance for the visual backbone is indirect, and the inconsistent issue still exists. To this end, we further propose UniQRNet, a multi-task framework over the QRNet to learn referring expression grounding and segmentation jointly. The framework introduces a multi-task head that leverages fine-grained pixel-level supervision from the segmentation task to directly guide the intermediate layers of QRNet to learn text-consistent visual features. Besides, UniQRNet also includes a loss balance strategy that allows two types of supervision signals to cooperate and optimize the model together. We conduct the most comprehensive comparison experiment covering four major datasets, ten evaluation set and three evaluation metrics used in previous work. UniQRNet outperforms previous state-of-the-art methods by a large margin on both referring comprehensive grounding (1.8%~5.09%) and segmentation tasks (0.57%~5.56%). Ablation and analysis reveal that UniQRNet can improve the consistency of visual features with text input and can bring significant performance improvement.

BarlowTwins-CXR: enhancing chest X-ray abnormality localization in heterogeneous data with cross-domain self-supervised learning.

Chest X-ray imaging based abnormality localization, essential in diagnosing various diseases, faces significant clinical challenges due to complex interpretations and the growing workload of radiologists. While recent advances in deep learning offer promising solutions, there is still a critical issue of domain inconsistency in cross-domain transfer learning, which hampers the efficiency and accuracy of diagnostic processes. This study aims to address the domain inconsistency problem and improve autonomic abnormality localization performance of heterogeneous chest X-ray image analysis, particularly in detecting abnormalities, by developing a self-supervised learning strategy called "BarlwoTwins-CXR". We utilized two publicly available datasets: the NIH Chest X-ray Dataset and the VinDr-CXR. The BarlowTwins-CXR approach was conducted in a two-stage training process. Initially, self-supervised pre-training was performed using an adjusted Barlow Twins algorithm on the NIH dataset with a Resnet50 backbone pre-trained on ImageNet. This was followed by supervised fine-tuning on the VinDr-CXR dataset using Faster R-CNN with Feature Pyramid Network (FPN). The study employed mean Average Precision (mAP) at an Intersection over Union (IoU) of 50% and Area Under the Curve (AUC) for performance evaluation. Our experiments showed a significant improvement in model performance with BarlowTwins-CXR. The approach achieved a 3% increase in mAP50 accuracy compared to traditional ImageNet pre-trained models. In addition, the Ablation CAM method revealed enhanced precision in localizing chest abnormalities. The study involved 112,120 images from the NIH dataset and 18,000 images from the VinDr-CXR dataset, indicating robust training and testing samples. BarlowTwins-CXR significantly enhances the efficiency and accuracy of chest X-ray image-based abnormality localization, outperforming traditional transfer learning methods and effectively overcoming domain inconsistency in cross-domain scenarios. Our experiment results demonstrate the potential of using self-supervised learning to improve the generalizability of models in medical settings with limited amounts of heterogeneous data. This approach can be instrumental in aiding radiologists, particularly in high-workload environments, offering a promising direction for future AI-driven healthcare solutions.

A New Test Toward Consistency in International Investment Arbitration

Abstract The issue of inconsistency in international investment arbitration has become more evident with the proliferation of investment disputes. This article analyzes the value as well as the desired extent of consistency and coherence for investment arbitration, and further proposes a new test that would enhance consistency and foster the development of investment law. By matching the current status of investment arbitration with three identified conditions for the application of stare decisis or jurisprudence constante, this article concludes that strictly applying the doctrines to achieve consistency is neither feasible nor favorable for the investment arbitration regime. Further, while the public nature of this dispute-resolution mechanism requires consistency and coherence, existing attempts and approaches in this regard have significant limitations. As such, this article proposes a soft precedents test to reach a balance between consistency and flexibility. The test requires the tribunals to consider previous cases in good faith when facing the same or similar issues discussed therein, and the core consideration in whether to agree with previous cases is to develop international investment law. It is suggested that this proposed test would help to clarify, enrich and develop the content of the law within the investment arbitration community.

Addressing inconsistency in use of the LMS

Inconsistency in the use of the learning management system (LMS) by academic staff is a source of dissatisfaction among university students in the UK. One solution is to establish a set of minimum standards (or baseline) for LMS use within an individual institution. Another is to supply templates – frameworks for LMS course sites – with a view to providing students with a seamless experience in their interactions with the LMS. This paper describes how the issue of inconsistency was addressed at a leading research university in the UK through an exploratory project, WebLearn Improved Student Experience (WISE). The widespread devolution of responsibility for site management to administrative staff, together with the ‘maverick’ creation of course sites by those academics who chose to engage with the WebLearn LMS, had resulted in unevenness in students’ access to learning materials. The project team engaged in close collaboration with 19 departments in order to achieve the immediate purpose of improving uptake of, and consistency in, their LMS presence. The ultimate aim was to develop a support package comprising LMS templates and ‘best practice’ guidelines that would enable departments in the future to achieve the same objective, either unsupported or with minimal assistance from the central team of learning technologists. The project was evaluated using a modification of the Innovation Histories method, which included interviews with 13 participants. The evaluation findings additionally threw into relief the complex social and cultural factors at play that can inhibit a consistent student experience in an institutional LMS.

Shopping trip recommendations: A novel deep learning-enhanced global planning approach

Brick-and-mortar shopping malls are embracing Artificial Intelligence (AI) technology and recommender systems to enhance the shopping experience and boost mall revenue. Echoing this trend, we formulate a new shopping trip recommendation problem, which aims to recommend a shopping trip (i.e., a list of stores) that matches customer preferences and has appropriate trip lengths. To solve this problem, we develop a novel deep learning-enhanced global planning (DeepGP) approach featuring three methodological novelties. First, we introduce a new shopping intensity term based on deep neural networks to capture the variation of trip lengths specific to different shopping contexts. Second, we innovatively formulate the learning and optimization objectives in a consistent form by balancing the shopping choice likelihood and the shopping intensity likelihood, thus resolving the inconsistency issue encountered by prior global planning methods. Third, to overcome the computational challenge caused by the nonlinear shopping intensity term, we design a new exact and efficient solution technique based on piecewise linear transformations. Using a real-world offline shopping dataset, we empirically demonstrate the superior performances of our approach compared to representative benchmarks in offering more accurate and relevant shopping trip recommendations. Through a simulation, we show the capacity of our approach to attract and balance customer traffic in practical deployments. Overall, our research highlights the efficacy of combining shopping choices and shopping intensity in a consistent learning and optimization framework for offline shopping trip recommendations.

Colour Image Enhancement Model of Retinal Fundus Image for Diabetic Retinopathy Recognition

Diabetic retinopathy (DR) features are typically identified through ophthalmologist eye examinations, but these images often facechallenges like low contrast, non-uniform illumination, and colour inconsistency, affecting the diagnosis accuracy. Therefore, this studyintroduces two novel techniques to improve image quality. One is applying colour image processing techniques to original retinalfundus images, overcoming existing algorithm limitations. Firstly, a new colour correction algorithm was proposed based on TunedBrightness Controlled Single-Scale Retinex (TBCSSR) named Fuzzy TBCSSR Histogram Matching (fTBCSSRhm) to address the issue of colour inconsistency in the dataset. Secondly, based on hybrid particle swarm optimisation-contrast stretch (HPSOCS), the hybridof TBCSSR and HPSOCS named eTBCSSR-HPSOCS algorithm is introduced to tackle the limitations of the standard Particle SwarmOptimisation (PSO) algorithm in HPSOCS, which is prone to local optima and exhibits low convergence rates. This technique combinesthe L-component of the LAB colour model with an enhanced velocity mechanism in PSO and contrast stretching (lavHPSOCS). Its goalis to fine-tune parameters automatically, reduce over-enhancement, avoid unwanted artefacts, and preserve intricate details. This approach improves optimisation by balancing exploration and exploitation and refining velocity control. The proposed algorithm underwent both qualitative and quantitative evaluations. Tests on 100 retinal fundus images from primary datasets were performed to benchmark the algorithm against three existing approaches. The results show that the qualitative performance of the proposed enhancement is more favourable to ophthalmologist specialists than other images. Quantitatively, eTBCSSR-HPSOCS outperforms others with the lowest mean squared error (MSE) of 42.72859, the highest peak signal-to-noise ratio (PSNR) of 32.768, and entropy of 0.977.

Learning‐based monocular visual‐inertial odometry with SE2(3) $S{E}_{2}(3)$‐EKF

AbstractLearning‐based visual odometry (VO) becomes popular as it achieves a remarkable performance without manually crafted image processing and burdensome calibration. Meanwhile, the inertial navigation can provide a localization solution to assist VO when the VO produces poor state estimation under challenging visual conditions. Therefore, the combination of learning‐based technique and classical state estimation method can further improve the performance of pose estimation. In this paper, we propose a learning‐based visual‐inertial odometry (VIO) algorithm, which consists of an end‐to‐end VO network and an ‐Extended Kalman Filter (EKF). The VO network mainly combines a convolutional neural network with a recurrent neural network, taking advantage of two consecutive monocular images to produce relative pose estimation with associated uncertainties. The ‐EKF, which is proved to overcome the inconsistency issues of VIO, propagates inertial measurement unit kinematics‐based states, and fuses relative measurements and uncertainties from the VO network in its update step. The extensive experimental results on the KITTI data set and the EuRoC data set demonstrate the superior performance of the proposed method compared to other related methods.

Robust condition identification against label noise in industrial processes based on trusted connection dictionary learning

In the era of big data, the pervasive use of artificial intelligence (AI) technology has revolutionized various industries. AI-powered systems, particularly those utilizing data-driven interpretable and reliable methods for accurate condition identification, have emerged as a crucial factor in achieving intelligent industrial systems. However, in real industrial sites, the labels of offline data often contain incorrect labels, i.e. label noise, due to human error, sensor failure, and other sources, which reduces the accuracy of the model and affects the performance and safety of the industrial system. To tackle the challenge of precarious performance caused by label noise, this paper proposes a robust method called Trusted Connection Dictionary Learning (TCDL) which effectively mitigates the adverse effects of label noise through interpretable constraints and guarantees the trustworthiness of the model for industrial applications. Specifically, to address the issue of data structure and label inconsistency arising from label noise, this paper presents a dynamic graph model that jointly optimizes data features and label information to obtain consistent representations of them. Subsequently, to effectively fit the labeled data, TCDL leverages the label information from the nearest neighbors through the optimized graph structure and facilitates the self-correction of label noise, leading to robust representations of process features and a reliable condition classifier. Finally, a novel iterative optimization algorithm that combines the Block Coordinate Descent (BCD) method and the Alternating Direction Multiplier Method (ADMM) is also developed to ensure the convergence of the algorithm. Due to simple constraints, it can obtain accurately optimized variables with few iterations. Numerical simulations and industrial experiments demonstrate the superiority of the proposed method compared to some state-of-the-art methods.

Onto-SAGCN: Ontology modeling and spatial attention-based graph convolution networks for aircraft assembly quality prediction

Aircraft assembly is an essential stage in the aircraft manufacturing industry, and the increasing complexity of aircraft functionality has put higher requirements for assembly quality. Various deep learning methods based on image or structural data have been used to predict assembly quality. However, these methods focus more on the structural relationships and interactions between assembled products and are difficult to adapt to scenarios where products and equipment may change. This paper proposes a graph convolutional neural network based on ontology modeling and spatial attention mechanism (Onto-SAGCN) for assembly quality analysis and prediction. Specifically, formal representations of assembly equipment, products, and assembly processes are established using ontologies, enabling the unified modeling of spatial and temporal relationships between entities and resolving the issues of data representation inconsistency. Then, the ontologies are transformed into undirected graphs, where nodes represent entities, node attributes represent feature data, and edges represent relationships or constraints between nodes. Designed to process data from the assembly process, the Onto-SAGCN model is subsequently applied to the assembly of aircraft moving wings. It predicts the diameters of holes by gathering operational data from equipment during the assembly and benchmarks these predictions against other established methods. Experimental outcomes affirm the method's elevated accuracy and dependability in forecasting assembly quality.

Temporally and Distributionally Robust Optimization for Cold-Start Recommendation

Collaborative Filtering (CF) recommender models highly depend on user-item interactions to learn CF representations, thus falling short of recommending cold-start items. To address this issue, prior studies mainly introduce item features (e.g., thumbnails) for cold-start item recommendation. They learn a feature extractor on warm-start items to align feature representations with interactions, and then leverage the feature extractor to extract the feature representations of cold-start items for interaction prediction. Unfortunately, the features of cold-start items, especially the popular ones, tend to diverge from those of warm-start ones due to temporal feature shifts, preventing the feature extractor from accurately learning feature representations of cold-start items. To alleviate the impact of temporal feature shifts, we consider using Distributionally Robust Optimization (DRO) to enhance the generation ability of the feature extractor. Nonetheless, existing DRO methods face an inconsistency issue: the worse-case warm-start items emphasized during DRO training might not align well with the cold-start item distribution. To capture the temporal feature shifts and combat this inconsistency issue, we propose a novel temporal DRO with new optimization objectives, namely, 1) to integrate a worst-case factor to improve the worst-case performance, and 2) to devise a shifting factor to capture the shifting trend of item features and enhance the optimization of the potentially popular groups in cold-start items. Substantial experiments on three real-world datasets validate the superiority of our temporal DRO in enhancing the generalization ability of cold-start recommender models.

DrFuse: Learning Disentangled Representation for Clinical Multi-Modal Fusion with Missing Modality and Modal Inconsistency

The combination of electronic health records (EHR) and medical images is crucial for clinicians in making diagnoses and forecasting prognoses. Strategically fusing these two data modalities has great potential to improve the accuracy of machine learning models in clinical prediction tasks. However, the asynchronous and complementary nature of EHR and medical images presents unique challenges. Missing modalities due to clinical and administrative factors are inevitable in practice, and the significance of each data modality varies depending on the patient and the prediction target, resulting in inconsistent predictions and suboptimal model performance. To address these challenges, we propose DrFuse to achieve effective clinical multi-modal fusion. It tackles the missing modality issue by disentangling the features shared across modalities and those unique within each modality. Furthermore, we address the modal inconsistency issue via a disease-wise attention layer that produces the patient- and disease-wise weighting for each modality to make the final prediction. We validate the proposed method using real-world large-scale datasets, MIMIC-IV and MIMIC-CXR. Experimental results show that the proposed method significantly outperforms the state-of-the-art models.

Multi-view orientational attention network combining point-based affinity for polyp segmentation

Most existing deep learning-based polyp segmentation methods neglect two important aspects of polyps: the geometric orientation information of polyps and the point information of the entire colonoscopy area. In this paper, we introduce a multi-view orientational attention network (MVOA-Net), which incorporates orientation and point awareness to effectively address the issue of intra-class inconsistency resulting from variations in polyp shape, size, and position, as well as the inter-class indistinction caused by the high similarity between polyp lesions and surrounding tissues. To achieve robust orientation awareness, we propose a novel geometric orientation transformer encoder (GOTE) based on horizontal and vertical views. Moreover, To simultaneously capture the global context information of GOTE and emphasize the important local information of the convolution-based attention encoder (CBAE), a global and local cross attention fusion module (CAFM) is also proposed to simultaneously model the long-range dependencies of polyps and pay sufficient attention to the local boundaries of polyps. Additionally, a efficient atrous spatial pyramid pooling (E-ASPP) module is proposed to enhance the semantic representation of high-level features. Finally, a point-based affinity module (PBAM) and a multi-scale fusion module (MSFM) are proposed to distinguish the disguise of polyps, further alleviating inter-class indistinction. The ablation study results demonstrate the effectiveness of each component. Quantitative and qualitative experimental results show that MVOA-Net achieves the best segmentation accuracy across domain polyp datasets and has obvious advantages in segmenting multiple polyp objects.

Advanced Persistent Threat Group Correlation Analysis via Attack Behavior Patterns and Rough Sets

In recent years, advanced persistent threat (APT) attacks have become a significant network security threat due to their concealment and persistence. Correlation analysis of APT groups is vital for understanding the global network security landscape and accurately attributing threats. Current studies on threat attribution rely on experts or advanced technology to identify evidence linking attack incidents to known APT groups. However, there is a lack of research focused on automatically discovering potential correlations between APT groups. This paper proposes a method using attack behavior patterns and rough set theory to quantify APT group relevance. It extracts two types of features from threat intelligence: APT attack objects and behavior features. To address the issues of inconsistency and limitations in threat intelligence, this method uses rough set theory to model APT group behavior and designs a link prediction method to infer correlations among APT groups. Experimental results on publicly available APT analysis reports show a correlation precision of 90.90%. The similarity coefficient accurately reflects the correlation strength, validating the method’s efficacy and accuracy.