Data Pruning Research Articles

Fish feeding behavior quantification and automatic acoustic event detection: algorithmic optimization with deep learning

In recent years, the integration of sensor processing technologies with intelligent feeding control systems has become crucial for enhancing recirculating aquaculture systems (RAS). Intelligent feeding control significantly improves profitability and animal welfare in aquaculture by accurately monitoring feed intake and reducing waste. Acoustic methods have proven effective in automating feed monitoring, particularly in assessing feeding intensity. Automatic Audio Event Detection (AED) holds promise for identifying sound events associated with feeding activities, enabling precise monitoring of feeding behavior and estimation of feed consumption through frequency-based feeding signatures. Advancements in sensor technologies and computational capabilities enable real-time collection, processing, and analysis of acoustic data. However, handling high volumes of data with minimal latency and high accuracy poses challenges. To address these challenges, we propose a systematic approach: (1) Data preprocessing using wavelet transforms enhances efficiency and reduces computational overhead. (2) Data pruning focuses on informative features and reduces input data dimensionality, improving classification accuracy. (3) We adopt a weakly-labeled semi-supervised learning model with Convolutional Neural Networks (CNNs) to detect feeding events, improving classification performance. (4) Ensemble learning combines multiple classifiers to mitigate overfitting and enhance model generalization. Our methodology shows promise in accurately quantifying fish feeding behavior using acoustic signals.

Thorough Data Pruning for Join Query in Database System

The improvement of robustness and efficiency for multi-way equijoin query is challenging, no-matter for centralized database systems or distributed database systems. Due to lots of unnecessary data existing during query processing, these two metrics will be seriously reduced. If we can thoroughly prune unnecessary data in advance, the robustness and efficiency will be highly improved. However, the pruning power of current strategies, such as predicate push-down and algebraic equivalence, is limited. We present deepDP, a powerful, generalized, and efficient strategy for data pruning. deepDP builds multiple independent pruning spaces by generating longest transitive closures and applies appropriate data pruning strategy for each pruning space. For thoroughly pruning unnecessary data, deepDP employs <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\alpha \cdot \beta$</tex-math></inline-formula> pruning strategy to clean each pruning space based on a newly designed statistic information-Hollow Range and re-shuffles the elements in all pruned spaces for maximizing robustness and efficiency benefits meanwhile minimizing the invasion. We implement deepDP in PostgreSQL but are not limited to it, and evaluate deepDP on TPC-H, JOB, and our synthesis benchmark–DHR. The experiment results show that compared to traditional data pruning strategy, deepDP can improve multi-way equijoin query on efficiency by 3.5x.

Application of RCS and signal-free RCS to tree-ring width and maximum latewood density data

Dendroclimatic research faces the challenge of selecting appropriate detrending methods for retaining low-frequency signals in temperature reconstructions. Among the numerous methods available to dendrochronologists, regional curve standardisation (RCS) and the signal-free approach in combination with RCS (SF-RCS) are increasingly used to preserve the full spectrum of temperature variance in tree-ring data. Here, we apply RCS and SF-RCS to tree-ring width (TRW) and maximum latewood density (MXD) datasets composed of only living and combined living and relict trees from northern Scandinavia. Whereas RCS and SF-RCS produce highly similar chronologies when applied to composite (living-plus-relict) datasets, particularly for MXD, both methods fail to establish chronologies coherent with regional temperature trends when applied to living-tree datasets. Additional tests including pruning of well-replicated living-tree datasets, to approximate the heterogenous age-structure of composite datasets, reveal improved results and coherent trends in MXD. While this demonstrates the applicability of joint detrending and pruning techniques to retain meaningful low-frequency variance in living-tree MXD chronologies, similar improvements were not achieved with TRW, likely because of the much stronger age-trend inherent to this widely used proxy. Further tests with other tree species and in alpine environments are needed to verify these findings. However, such assessments require an adjustment of tree-ring sampling protocols to increase replication to 50+ trees per site including old and young individuals to facilitate data pruning.

Utilizing Deep Learning Techniques to Identify People by Palm Print

Person recognition systems have been applied for several years, as fingerprint recognition has been experimented with different image resolutions for 15 years. Fingerprint recognition and biometrics for security are becoming commonplace. Biometric systems are emerging and evolving topics seen as fertile ground for researchers to investigate more deeply and discover new approaches. Among the most prominent of these systems is the palm printing system, which identifies individuals based on the palm of their hands because of the advantages that the palm possesses that cannot be replicated among humans, as in its theory of other fingerprints. This paper proposes a biometric system to identify people by handprint, especially palm area, using deep learning technology via a pre-trained model on the PolyU-IITD dataset. The proposed system goes through several basic stages, namely data pruning, processing, training, and prediction, and the results were promising, as the system's accuracy reached 90% based on the confusion matrix measures.

Genetic Susceptibility Profile of the Southern Indian Population of Hyderabad for Coronary Artery Disease

In order to portray the genetic susceptibility profile of the population of Hyderabad a prioritized set of 192 SNPs belonging to the 11q23.3, 9p21.3 and other CAD specific genomic regions were genotyped in a sample of ~500 cases and 500 controls. After data pruning, pair-wise logistic regression analysis of 148 CAD specific SNPs revealed significant association of 18 SNPs. The risk score analysis based on 18 SNPs and the ROC curve yielded an AUC of 0.822 suggesting substantial power of these SNPs to serve as predictive markers for CAD. Comparative analysis revealed certain SNPs belonging to APOA4-APOC3, APOA5-APOA4, BUD13, ZPR1, LPA and SMEK1 genes were associated significantly with the three phenotypic categories of CAD. The current study identified a unique set of SNPs to be associated with CAD and also portrays the distinct patterns of association of certain SNPs with the three major phenotypic sub categories of CAD in the population of Hyderabad.

Short Communication: Reduced GBLUP equations to core animals in the algorithm for proven and young (APY)

The number of animal genotypes is rapidly increasing, and a major challenge for animal models is inverting the genomic relationship matrix (G). Matrix G has a limited dimensionality, and the algorithm for proven and young (APY) makes inverting a large G possible via the inverse of a block diagonal of G with a size equivalent to the dimensionality of G. APY divides genotyped animals into core and non-core groups, and breeding values of non-core animals are conditioned on the breeding values of core animals. Therefore, there is the possibility of opting out equations for non-core animals from the model. A methodology was presented for a reduced APY genomic BLUP (GBLUP) to equations for core animals. Using a small example dataset, the method was validated by the equality of the full and the reduced model analysis results. Absorption of fixed effect equations into random effect equations was successful in reducing the number of equations to solve and producing the same random effect solutions. Extending the method to APY single-step GBLUP (ssGBLUP) was not computationally justifiable. Other reduction techniques exist for ssGBLUP (regardless of APY or non-APY) that work by reducing the number of equations for non-genotyped animals. The number of equations can further be reduced by data pruning.

Efficient Data Transfer Mechanism for DLMS/COSEM Enabled Smart Energy Metering Platform

We report our implementation of DLMS/COSEM (Device Language Message Specification/Companion Specification for Energy Metering) enabled advanced smart metering data collection using our in-house developed data-driven multivariate data pruning technique. Aiming at near real-time smart metering data collection, a working model of a smart energy meter with a data pruning subsystem is developed and incorporated with the data pruning and communication module. The implementation of DLMS/COSEM service layer enables standard pull-based data collection from multiple meters that are connected with a commercial head-end system. Our implementation enables to fetch smart meter readings from multiple meters with significantly reduced data footprint. Our empirical results demonstrate that, dynamic data pruning capability reduces the number of samples by nearly 85%. The bandwidth saving increases with the increasing batch size, by accounting for all protocol overheads the endto- end communication bandwidth saving is up to 85%. The DLMS overhead is negligible for larger batch size.

Attacking Deep Reinforcement Learning With Decoupled Adversarial Policy

While Deep Reinforcement Learning (DRL) has achieved outstanding performance in extensive applications, exploiting its vulnerability with adversarial attacks is essential towards building robust DRL systems. In this work, we aim to propose a novel Decoupled Adversarial Policy (DAP) for attacking the DRL mechanism, whereas the adversarial agent can decompose the adversarial policy into two separate sub-policies: 1) the switch policy which determines if an attacker should launch the attack, and 2) the lure policy which determines the action an attacker induces the victim to take. If the adversarial agent samples an injection action from the switch policy, the attacker can query the pre-constructed database for universal perturbation in the real-time manner, misleading the victim to take the induced action sampled from the lure policy. To train the adversarial agent to learn DAP, we utilize those samples wherein both of the sub-actions from DAP are not restricted by each other or by the external constraint, but can actually affect the attacker’s behaviors. Therefore, we propose trajectory clipping and padding in data pruning, and Decoupled Proximal Policy Optimization (DPPO) in optimizing. Extensive experiments on different Atari games demonstrate the effectiveness of our proposed method. In addition, it can simultaneously implement the real-time and few-steps attack, which outperforms the existing counterparts.

On a black hole effect in bilinear curve resolution based on least squares

AbstractLeast squares‐based estimations lay behind most chemometric methodologies. Their properties, though, have been extensively studied mainly in the domain of regression, in relation to which the effect of well‐known deleterious factors (like object leverage or data distributions deviating from ideal conditions) on the accuracy of the prediction of an external response variable has been thoroughly assessed. Conversely, much less attention has been paid to what these factors might yield in alternative scenarios, where least squares approaches are still utilised, yet the objectives of data modelling may be very different. As an example, one can think of multivariate curve resolution (MCR) problems which are usually addressed by means of multivariate curve resolution‐alternating least squares (MCR‐ALS). In this respect, this article wants to offer a perspective on the basic principles of MCR‐ALS from the regression point of view. In particular, the following critical aspects will be highlighted: (i) in the presence of minor components, if the number of analysed data points is too large, the leverage of those that may be essential for a MCR‐ALS resolution might become too low for guaranteeing its correctness, and (ii) in order to overcome this black hole effect and improve the accuracy of the MCR‐ALS output, data pruning can be exploited. More in detail, this communication will provide a practical illustration of such aspects in the field of hyperspectral imaging where even single experimental runs may lead to the generation of massive amounts of spectral recordings.

Multimodal data augmentation for digital twining assisted by artificial intelligence in mechanics of materials

Digital twins in the mechanics of materials usually involve multimodal data in the sense that an instance of a mechanical component has both experimental and simulated data. These simulations aim not only to replicate experimental observations but also to extend the data. Whether spatially, temporally, or functionally, augmentation is needed for various possible uses of the components to improve the predictions of mechanical behavior. Related multimodal data are scarce, high-dimensional and a physics-based causality relation exists between observational and simulated data. We propose a data augmentation scheme coupled with data pruning, in order to limit memory requirements for high-dimensional augmented data. This augmentation is desirable for digital twining assisted by artificial intelligence when performing nonlinear model reduction. Here, data augmentation aims at preserving similarities in terms of the validity domain of reduced digital twins. In this article, we consider a specimen subjected to a mechanical test at high temperature, where the as-manufactured geometry may impact the lifetime of the component. Hence, an instance is represented by a digital twin that includes 3D X-Ray tomography data of the specimen, the related finite element mesh, and the finite element predictions of thermo-mechanical variables at several time steps. There is, thus, for each specimen, geometrical and mechanical information. Multimodal data, which couple different representation modalities together, are hard to collect, and annotating them requires a significant effort. Thus, the analysis of multimodal data generally suffers from the problem of data scarcity. The proposed data augmentation scheme aims at training a recommending system that recognizes a category of data available in a training set that has already been fully analyzed by using high-fidelity models. Such a recommending system enables the use of a ROM-net for fast lifetime assessment via local reduced-order models.

The Mann-Kendall-Sneyers test to identify the change points of COVID-19 time series in the United States

BackgroundOne critical variable in the time series analysis is the change point, which is the point where an abrupt change occurs in chronologically ordered observations. Existing parametric models for change point detection, such as the linear regression model and the Bayesian model, require that observations are normally distributed and that the trend line cannot have extreme variability. To overcome the limitations of the parametric model, we apply a nonparametric method, the Mann-Kendall-Sneyers (MKS) test, to change point detection for the state-level COVID-19 case time series data of the United States in the early outbreak of the pandemic.MethodsThe MKS test is implemented for change point detection. The forward sequence and the backward sequence are calculated based on the new weekly cases between March 22, 2020 and January 31, 2021 for each of the 50 states. Points of intersection between the two sequences falling within the 95% confidence intervals are identified as the change points. The results are compared with two other change point detection methods, the pruned exact linear time (PELT) method and the regression-based method. Also, an open-access tool by Microsoft Excel is developed to facilitate the model implementation.ResultsBy applying the MKS test to COVID-19 cases in the United States, we have identified that 30 states (60.0%) have at least one change point within the 95% confidence intervals. Of these states, 26 states have one change point, 4 states (i.e., LA, OH, VA, and WA) have two change points, and one state (GA) has three change points. Additionally, most downward changes appear in the Northeastern states (e.g., CT, MA, NJ, NY) at the first development stage (March 23 through May 31, 2020); most upward changes appear in the Western states (e.g., AZ, CA, CO, NM, WA, WY) and the Midwestern states (e.g., IL, IN, MI, MN, OH, WI) at the third development stage (November 19, 2020 through January 31, 2021).ConclusionsThis study is among the first to explore the potential of the MKS test applied for change point detection of COVID-19 cases. The MKS test is characterized by several advantages, including high computational efficiency, easy implementation, the ability to identify the change of direction, and no assumption for data distribution. However, due to its conservative nature in change point detection and moderate agreement with other methods, we recommend using the MKS test primarily for initial pattern identification and data pruning, especially in large data. With modification, the method can be further applied to other health data, such as injuries, disabilities, and mortalities.

Input reduction of convolutional neural networks with global sensitivity analysis as a data-centric approach

Pruning methods are used for dealing with the rapid growth of neural network parameters as the neural network develops. These enable a reduction in not only the size of the network, but also the bandwidth it utilizes. In this article, global sensitivity analysis methods, like Sobol and eFAST, are applied to determine the least significant inputs. Data-centric Artificial Intelligence is a fledgeling science and is gathering more and more attention among top researchers. This methodology is focused on data manipulation and constant model architecture, while the model-centric approach modifies models. This article cuts off itself from typical model-centric pruning methods and is focused solely on data pruning. In the work, principal component analysis is applied as a benchmark for the proposed Global Sensitivity methods. This work is focused on convolutional neural networks. It shows how the accuracy of CNNs is impacted by input reduction – doing so through the example of three image datasets and three series datasets.

Deep learning-based breast cancer disease prediction framework for medical industries

Breast cancer is one among the dreadful cancer which is caused due to formation in breast cells. Earlier recognition of breast cancer is most required in the medical field to avoid the dangerous threat to human life. This is carried out in the existing work, namely Predictive Modeling Technique (PMT). Existing work cannot handle the database with noises properly which might lead to inaccurate prediction outcome. These problems are addressed by introducing Deep Learning-based Breast Cancer Disease Prediction Framework (DLBCDPF). The proposed research framework objective is to present the structures for the disease diagnosis in a further accurate way. In this work, feature selection is achieved through optimization algorithm, namely ranking-based bee colony approach by which the most optimal feature is chosen from the training dataset. The fitness values considered in this work for optimal feature selection are F-score values. Each feature’s F-score and N numbers of feature’s F-score are arranged in a descending manner; in addition, feature subset with more than one feature are produced. In this phase, diagnosis of various stomach-related problems is done through a unique hybridized classification methodology. In this hybridization methodology, clustering is accomplished before classification, and data pruning is attained in every classification iteration. This leads to improved classification accuracy owing to efficient diagnosis. The clustering is attained by fuzzy C-means clustering, and classification is done using the improved deep neural network. The entire research analysis is carried out in python platform for breast cancer dataset from which it is substantiated that the suggested research work tends to outperform in an enhanced way than prevailing work.

Analysis of 61 SNPs from the CAD specific genomic loci reveals unique set of SNPs as significant markers in the Southern Indian population of Hyderabad

BackgroundThe present study is a part of the major project on coronary artery disease (CAD) carried out at Indian Statistical Institute, Hyderabad to investigate the pattern of association of SNPs selected from the CAD specific genomic loci. The study is expected to portray the genetic susceptibility profile of CAD specifically in the Southern Indian population of Hyderabad.MethodsThe study was conducted in a cohort of 830 subjects comprising 350 CAD cases and 480 controls from Hyderabad. A prioritized set of 61 SNPs selected from the NHGRI GWAS catalogue were genotyped using FluidigmNanofluidic SNP Genotyping System and appropriate statistical analyses were used in interpreting the results.ResultsAfter data pruning, out of 45 SNPs qualified for the association analysis, four SNPs were found to be highly significantly associated with increased risk for CAD even after Bonferroni correction for multiple testing (p < 0.001). These results were also replicated in the random subsets of the pooled cohort (70, 50 and 30%) suggesting internal consistency. The ROC analysis of the risk scores of the significant SNPs suggested highly significant area under curve (AUC = 0.749; p < 0.0001) implying predictive utility of these risk variants.ConclusionsThe rs10455872 of LP(A) gene in particular showed profound risk for CAD (OR 35.9; CI 16.7–77.2) in this regional Indian population. The other significant SNP associations observed with respect to the pooled CAD cohort and in different anatomical and phenotypic severity categories reflected on the role of genetic heterogeneity in the clinical heterogeneity of CAD. The SNP rs7582720 of WDR12 gene, albeit not individually associated with CAD, was found to be conferring significant risk through epistatic interaction with two SNPs (rs6589566, rs1263163 in ZPR1, APOA5-APOA4 genes) of the 11q23.3 region.

An Edge Computing-oriented Net Power Forecasting for PV-assisted Charging Station: Model Complexity and Forecasting Accuracy Trade-off

The PV-assisted charging station (PVCS) aggregates the two key resources of electric vehicle (EV) charging load and photovoltaic (PV) system to maximize operation profit. In order to quantify the PVCS impact to grid and the renewable energy utilization by EV charging load, it is crucial to predict the PVCS net power and capture the most correlated factors for power variation. Given that the PVCS is located near the user side, the forecasting model complexity must be restricted to meet the online training demands on edge computing, apart from ensuring the prediction accuracy. A PVCS net power forecasting approach is proposed by simplifying model complexity from the entire cycle of training process, including the input data pruning, lightweight model training and hyperparameter optimization. Finally, a comprehensive analysis of the real data in PVCS shows that the most sensitivity factors affecting the net power of PVCS is time of use price, and the deep auto-encoded extreme learning machine (DA-ELM) can make a satisfactory compromise between prediction accuracy and model complexity for edge computing utilization. The forecasting model has outstanding prediction performance on Raspberry pi-based edge platform, which has enough significance for PVCS promotion.

RecoMed: A knowledge-aware recommender system for hypertension medications

Background & ObjectiveHigh medicine diversity has always been a significant challenge for prescription, causing confusion or doubt in physicians’ decision-making process. This paper aims to develop a medicine recommender system called RecoMed to aid the physician in the prescription process of hypertension by providing information about what medications have been prescribed by other doctors and figuring out what other medicines can be recommended in addition to the one in question. MethodsThere are two steps to the developed method: First, association rule mining algorithms are employed to find medicine association rules. The second step entails graph mining and clustering to present an enriched recommendation via ATC code, which itself comprises several steps. First, the initial graph is constructed from historical prescription data. Then, data pruning is performed in the second step, after which the medicines with a high repetition rate are removed at the discretion of a general medical practitioner. Next, the medicines are matched to a well-known medicine classification system called the ATC code to provide an enriched recommendation. And finally, the DBSCAN and Louvain algorithms cluster medicines in the final step. ResultsA list of recommended medicines is provided as the system's output, and physicians can choose one or more of the medicines based on the patient's clinical symptoms. Only the medicines of class #2, related to high blood pressure medications, are used to assess the system's performance. Unlike other research studies, in this work, the framework proposed is entirely based on unsupervised learning methods and so it is more likely to be used in real-world applications as medical data are generally unlabeled, and supervised learning models cannot be directly built using these types of data. The results obtained from this system have been reviewed and confirmed by an expert in this field.

Efficient and Robust Online Trajectory Prediction for Non-Cooperative Unmanned Aerial Vehicles

As an important type of dynamic data-driven application system, unmanned aerial vehicles (UAVs) are widely used for civilian, commercial, and military applications across the globe. An increasing research effort has been devoted to trajectory prediction for non-cooperative UAVs to facilitate their collision avoidance and trajectory planning. Existing methods for UAV trajectory prediction typically suffer from two major drawbacks: inadequate uncertainty quantification of the impact of external factors (e.g., wind) and inability to perform online detection of abrupt flying pattern changes. This paper proposes a Gaussian process regression (GPR)-based trajectory prediction framework for UAVs featuring three novel components: 1) GPR with uniform confidence bounds for simultaneous predictive uncertainty quantification, 2) online trajectory change-point detection, and 3) adaptive training data pruning. The paper also demonstrates the superiority of the proposed framework to competing trajectory prediction methods via numerical studies using both simulation and real-world datasets.

Real-Time Stress Level Feedback from Raw Ecg Signals for Personalised, Context-Aware Applications Using Lightweight Convolutional Neural Network Architectures.

Human stress is intricately linked with mental processes such as decision making. Public protection practitioners, including Law Enforcement Agents (LEAs), are forced to make difficult decisions during high-pressure operations, under strenuous circumstances. In this respect, systems and applications that assist such practitioners to take decisions, are increasingly incorporating user stress level information for their development, adaptation, and evaluation. To that end, our goal is to accurately detect and classify the level of acute, short-term stress, in real time, for the development of personalized, context-aware solutions for LEAs. Deep Neural Networks (DNNs), and in particular Convolutional Neural Networks (CNNs), have been gaining traction in the field of stress analysis, exhibiting promising results. Furthermore, the electrocardiogram (ECG) signals, have also been widely adopted for estimating levels of stress. In this work, we propose two CNN architectures for the stress detection and 3-level (low, moderate, high) stress classification tasks, using ultra short-term raw ECG signals (3 s). One architecture is simple and with a low memory footprint, suitable for running in wearable edge-computing nodes, and the other is able to learn more complex features, having more trainable parameters. The models were trained on the two publicly available stress classification datasets, after applying pre-processing techniques, such as data pruning, down-sampling, and data augmentation, using a sliding window approach. After hyperparameter tuning, using 4-fold cross-validation, the evaluation on the test set demonstrated state-of-the-art accuracy both on the 3- and 2-level stress classification task using the DriveDB dataset, reporting an accuracy of 83.55% and 98.77% respectively.

PEDF, a pleiotropic WTC-LI biomarker: Machine learning biomarker identification and validation.

Biomarkers predict World Trade Center-Lung Injury (WTC-LI); however, there remains unaddressed multicollinearity in our serum cytokines, chemokines, and high-throughput platform datasets used to phenotype WTC-disease. To address this concern, we used automated, machine-learning, high-dimensional data pruning, and validated identified biomarkers. The parent cohort consisted of male, never-smoking firefighters with WTC-LI (FEV1, %Pred< lower limit of normal (LLN); n = 100) and controls (n = 127) and had their biomarkers assessed. Cases and controls (n = 15/group) underwent untargeted metabolomics, then feature selection performed on metabolites, cytokines, chemokines, and clinical data. Cytokines, chemokines, and clinical biomarkers were validated in the non-overlapping parent-cohort via binary logistic regression with 5-fold cross validation. Random forests of metabolites (n = 580), clinical biomarkers (n = 5), and previously assayed cytokines, chemokines (n = 106) identified that the top 5% of biomarkers important to class separation included pigment epithelium-derived factor (PEDF), macrophage derived chemokine (MDC), systolic blood pressure, macrophage inflammatory protein-4 (MIP-4), growth-regulated oncogene protein (GRO), monocyte chemoattractant protein-1 (MCP-1), apolipoprotein-AII (Apo-AII), cell membrane metabolites (sphingolipids, phospholipids), and branched-chain amino acids. Validated models via confounder-adjusted (age on 9/11, BMI, exposure, and pre-9/11 FEV1, %Pred) binary logistic regression had AUCROC [0.90(0.84–0.96)]. Decreased PEDF and MIP-4, and increased Apo-AII were associated with increased odds of WTC-LI. Increased GRO, MCP-1, and simultaneously decreased MDC were associated with decreased odds of WTC-LI. In conclusion, automated data pruning identified novel WTC-LI biomarkers; performance was validated in an independent cohort. One biomarker—PEDF, an antiangiogenic agent—is a novel, predictive biomarker of particulate-matter-related lung disease. Other biomarkers—GRO, MCP-1, MDC, MIP-4—reveal immune cell involvement in WTC-LI pathogenesis. Findings of our automated biomarker identification warrant further investigation into these potential pharmacotherapy targets.

Assessment of safety leadership among senior managers in China's coal mining industry

The factors giving rise to the safety leadership of senior managers in the coal mining industry are underexplored. We conducted 24 semistructured interviews with experienced senior managers in this industry. We used hermeneutic content analysis for the interview content, identified five safety leadership dimensions of senior mangers, and designed the Safety Leadership Scale in Coal Mines in China. Data from 845 participants in nine coal mines were analyzed with confirmatory factor analysis (CFA). We followed a two-step approach, involving data pruning and a second-order CFA, and validated a 21-item measure that shows good validity and reliability. The secondorder CFA results show that the measure yielded an acceptable fit. Our results can be used to estimate the safety leadership level among senior managers in China's coal mining industry.