Gaussian Model Research Articles

Emotion Regulation, Eating Psychopathology, and Putative Transdiagnostic Psychological Processes: Findings from an Exploratory Network Analysis in a College Sample

Objective: Considering the prevalence of ED-related prodromal symptoms among higher education students (making them a population at risk for developing EDs), the main goals of this study were to conduct a network analysis in a college sample and to explore multivariate dependencies between a selection of empirically informed variables of interest to eating psychopathology, namely difficulties in emotion regulation and psychological processes (e.g., interoceptive awareness, self-compassion, self-criticism, mindfulness, and experiential avoidance). Methods: The sample included 294 college students (Mage = 21.4, SD = 5.0; MBMI = 22.4, SD = 3.7). A Gaussian graphical network model was estimated to visualize interactions among the studied variables and to assess their centrality in terms of betweenness, closeness, strength, and expected influence. Results: A network system with 21 nodes was estimated (sparsity = 0.52). Nodes assessing disordered eating symptoms displayed the strongest correlation coefficients with nodes assessing dimensions of interoceptive awareness: eating concerns and not-distracting (r = −0.11), shape concerns and trusting (r = −0.16), and weight concerns and trusting (r = −0.10). Self-compassion was the node with the highest betweenness (SELFCS = 2.27) and closeness centrality (SELFCS = 1.70). The nodes with the highest strength centrality were strategies (DERS = 1.91) and shape concerns (EDE-Q = 1.51). Discussion: In this network model conducted in a college sample, eating-related symptoms were mainly associated with dimensions of interoceptive awareness. Also, the lack of effective strategies to regulate emotions, shape concerns, and self-compassion stood out as central nodes in the network model. The results suggest that addressing these variables may be promising in disrupting network systems marked by the presence of prodromal eating psychopathology symptoms in at-risk populations (e.g., college students).

Video SAR Moving Target Detection Method Based on Machine Vision

When detecting moving targets in video SAR, it is impossible to accurately obtain the moving targets in dynamic images. In order to solve the problems of low detection integrity and poor effectiveness, a video SAR moving target detection method based on machine vision is proposed. Firstly, the moving target pixel points are matched by the mixed Gaussian model to obtain the target area of the video SAR moving image; secondly, the HSV model and reflectivity algorithm are used to remove the shadows caused by the moving target; finally, the processed target area is input into the YOLO algorithm to complete the final video SAR moving target detection. Experimental results show that the proposed algorithm has high detection integrity, high detection rate, low false detection rate, and better detection effectiveness.

Stable and unstable fall motions of plate-like ice crystal analogues

Abstract. The orientation of ice crystals affects their microphysical behaviour, growth, and precipitation. Orientation also affects interaction with electromagnetic radiation, and through this it influences remote sensing signals, in situ observations, and optical effects. Fall behaviours of a variety of 3D-printed plate-like ice crystal analogues in a tank of water–glycerine mixture are observed with multi-view cameras and digitally reconstructed to simulate the falling of ice crystals in the atmosphere. Four main falling regimes were observed: stable, zigzag, transitional, and spiralling. Stable motion is characterised by no resolvable fluctuations in velocity or orientation, with the maximum dimension oriented horizontally. The zigzagging regime is characterised by a back-and-forth swing in a constant vertical plane, corresponding to a time series of inclination angle approximated by a rectified sine wave. In the spiralling regime, analogues consistently incline at an angle between 7 and 28°, depending on particle shape. Transitional behaviour exhibits motion in between spiral and zigzag, similar to that of a falling spherical pendulum. The inclination angles that unstable planar ice crystals make with the horizontal plane are found to have a non-zero mode. This observed behaviour does not fit the commonly used Gaussian model of inclination angle. The typical Reynolds number when oscillations start is strongly dependent on shape: solid hexagonal plates begin to oscillate at Re =237, whereas several dendritic shapes remain stable throughout all experiments, even at Re > 1000. These results should be considered within remote sensing applications wherein the orientation characteristics of ice crystals are used to retrieve their properties.

Gaussian Mixture Regression Model with Sparsity for Clustering of Territory Risk in Auto Insurance

Abstract Insurance rating territory design and accurate estimation of territory risk relativities are fundamental aspects of auto insurance rate regulation. It is crucial to develop methodologies that can facilitate the effective design of rating territories and their risk relativities estimate, as they directly impact the rate filing and the decision support of the rate change review process. This article proposes a Gaussian Mixture Regression model clustering approach for territory design. The proposed method incorporates a linear regression model, taking spatial location as model covariates, which helps estimate the cluster mean more accurately. Also, to further enhance the estimation of territory risk relativities, we impose sparsity through sparse matrix decomposition of the membership coefficient matrix obtained from the Gaussian Mixture Regression model. By transitioning from the current hard clustering method to a soft approach, our methodology could improve the evaluation of territory risk for rate-making purposes. Moreover, using non-negative sparse matrix approximation ensures that the estimation of risk relativities for basic rating units remains smooth, effectively eliminating data noise from the territory risk relativity estimate. Overall, our novel methodology aims to significantly enhance the accuracy and reliability of risk analysis in auto insurance. Furthermore, the proposed method exhibits potential for extension to various other domains that involve spatial clustering of data, thereby broadening its applicability and expanding its usefulness beyond auto insurance rate regulation.

Distance preserving machine learning for uncertainty aware accelerator capacitance predictions

Abstract Accurate uncertainty estimations are essential for producing reliable machine learning models, especially in safety-critical applications such as accelerator systems. Gaussian process models are generally regarded as the gold standard for this task; however, they can struggle with large, high-dimensional datasets. Combining deep neural networks with Gaussian process approximation techniques has shown promising results, but dimensionality reduction through standard deep neural network layers is not guaranteed to maintain the distance information necessary for Gaussian process models. We build on previous work by comparing the use of the singular value decomposition against a spectral-normalized dense layer as a feature extractor for a deep neural Gaussian process approximation model and apply it to a capacitance prediction problem for the High Voltage Converter Modulators in the Oak Ridge Spallation Neutron Source. Our model shows improved distance preservation and predicts in-distribution capacitance values with less than 1% error.

Analysis of mapping atomic models to coarse-grained resolution.

Low-resolution coarse-grained (CG) models provide significant computational and conceptual advantages for simulating soft materials. However, the properties of CG models depend quite sensitively upon the mapping, M, that maps each atomic configuration, r, to a CG configuration, R. In particular, M determines how the configurational information of the atomic model is partitioned between the mapped ensemble of CG configurations and the lost ensemble of atomic configurations that map to each R. In this work, we investigate how the mapping partitions the atomic configuration space into CG and intra-site components. We demonstrate that the corresponding coordinate transformation introduces a nontrivial Jacobian factor. This Jacobian factor defines a labeling entropy that corresponds to the uncertainty in the atoms that are associated with each CG site. Consequently, the labeling entropy effectively transfers configurational information from the lost ensemble into the mapped ensemble. Moreover, our analysis highlights the possibility of resonant mappings that separate the atomic potential into CG and intra-site contributions. We numerically illustrate these considerations with a Gaussian network model for the equilibrium fluctuations of actin. We demonstrate that the spectral quality, Q, provides a simple metric for identifying high quality representations for actin. Conversely, we find that neither maximizing nor minimizing the information content of the mapped ensemble results in high quality representations. However, if one accounts for the labeling uncertainty, Q(M) correlates quite well with the adjusted configurational information loss, Îmap(M), that results from the mapping.

Characterization of plugging and unplugging process for electric vehicle charging connectors based on Force/Torque measurements

Understanding the magnitude and profile of forces and torques (F/T) to properly execute the plugging and unplugging for electric vehicles (EVs) charging is crucial for designing reliable robots in automated charging stations (ACS). Previous research focuses on F/T measurement for controlling force during robotic tasks. In contrast, this study aims to refine automation workflows for smoother operation. By utilizing the comprehensive F/T measurements, it becomes possible to extract precise information during the plugging and unplugging process. This data provides insight into interaction dynamics, improving the robot’s ability to perform tasks more accurately and efficiently. Through experiments involving 40 participants, anthropometric and F/T data were analysed using statistics, uncovering key relationships between height, weight, and forces. The F/T profiles were predicted using polynomial, Fourier series, and Gaussian models, with the Gaussian proving the most accurate. Peak force and impulse are significantly higher during plugging (74.15 Ns) than unplugging(57.92Ns).

A cardiac audio classification method based on image expression of multidimensional features.

Heart sound auscultation plays a crucial role in the early diagnosis of cardiovascular diseases. In recent years, great achievements have been made in the automatic classification of heart sounds, but most methods are based on segmentation features and traditional classifiers and do not fully exploit existing deep networks. This paper proposes a cardiac audio classification method based on image expression of multidimensional features (CACIEMDF). First, a 102-dimensional feature vector is designed by combining the characteristics of heart sound data in the time domain, frequency domain and statistical domain. Based on the feature vector, a two-dimensional feature projection space is constructed by PCA dimensionality reduction and the convex hull algorithm, and 102 pairs of coordinate representations of the feature vector in the two-dimensional space are calculated. Each one-dimensional component of the feature vector corresponds to a pair of 2D coordinate representations. Finally, the one-dimensional feature component value and its divergence into categories are used to fill the three channels of a color image, and a Gaussian model is used to dye the image to enrich its content. The color image is sent to a deep network such as ResNet50 for classification. In this paper, three public heart sound datasets are fused, and experiments are conducted using the above methods. The results show that for the two-classification/five-classification task of heart sounds, the method in this paper can achieve a classification accuracy of 95.68%/94.53% when combined with the current deep network.

Gaussian distribution-based randomised image method

Sweeping echo is an acoustical phenomenon present in perfectly rectangular rooms that is characterised by a continuous increase in pitch arising from the concurrent arrival of multiple higher-order reflections. Even though the room acoustic simulation software predicts their presence, in most rectangular rooms sweeping echoes can hardly be perceived due to the existence of small irregularities in the room geometry. In this paper, the image sources' position randomisation model developed within the randomised image method framework with which the incongruence between the original image method results and in situ measurements can successfully be reduced is replaced with a Gaussian distribution model. The results indicate that with the proposed model an improvement in the stability of the room transfer function across realisations can be achieved.

A network analysis of affective and motivational individual differences and error monitoring in a non-clinical sample.

Error monitoring, which plays a crucial role in shaping adaptive behavior, is influenced by a complex interplay of affective and motivational factors. Understanding these associations often proves challenging due to the intricate nature of these variables. With the aim of addressing previous inconsistencies and methodological gaps, in this study, we utilized network analysis to investigate the relationship between affective and motivational individual differences and error monitoring. We employed six Gaussian Graphical Models on a non-clinical population ($N$ = 236) to examine the conditional dependence between the amplitude of response-related potentials (error-related negativity; correct-related negativity) and 29 self-report measures related to anxiety, depression, obsessive thoughts, compulsive behavior, and motivation while adjusting for covariates: age, handedness, and latency of error-related negativity and correct-related negativity. We then validated our results on an independent sample of 107 participants. Our findings revealed unique associations between error-related negativity amplitudes and specific traits. Notably, more pronounced error-related negativity amplitudes were associated with increased rumination and obsessing, and decreased reward sensitivity. Importantly, in our non-clinical sample, error-related negativity was not directly associated with trait anxiety. These results underscore the nuanced effects of affective and motivational traits on error processing in healthy population.

Health impact assessment of exposure to road traffic noise and air pollution according to pre- and post-densification scenarios in Helsingborg, Sweden

A prevalent solution to accommodate population growth due to urbanization is densification. However, this often pushes new residential housing closer to roads, increasing exposure to both noise and air pollution. The present study’s aim was to estimate the health impacts of road traffic-related noise and air pollution for a low-income area (Drottninghög) in Helsingborg, Sweden, according to pre-densification (2012) and post-densification (2030) scenarios.Road traffic noise was simulated at the façade of residential buildings using the Nordic prediction method, and exposure was assessed using SoundPLAN. Exposure-response functions (ERF) from the WHO were utilized for the following health outcomes associated with noise: annoyance, adverse sleep disturbance, ischemic heart disease (IHD) incidence and IHD mortality. Air pollution (nitrogen dioxide, NO2) was assessed using a Gaussian dispersion model (AERMODE). Health outcomes associated with NO2 included natural cause mortality, pediatric asthma, respiratory hospitalizations and low birth weight (LBW). ERFs were derived from meta-analyses. Health impact assessments were then performed for both scenarios.Densifying Drottninghög according to the municipality’s planned strategy would lead to a 15 % unit increase in the proportion of residents exposed to road traffic noise above the WHO’s health-based guideline value (53 dB(A) Lden). This was estimated to markedly increase the proportion of residents highly annoyed by traffic noise (7.4–13.9 %) as well as those highly sleep disturbed (3.0–4.9 %). IHD incidence and IHD mortality attributed to noise would increase by an estimated 49 % and 44 %, respectively, post-densification. NO2 exposure was estimated to increase slightly (0.7 µg/m3) post-densification, which would contribute to an estimated 4–6 % increase in natural cause mortality, pediatric asthma, respiratory hospitalization and LBW.Urban planning initiatives need to consider these prevalent urban environmental exposures and integrate a public health perspective into densification strategies. Doing so can create synergies in the built environment that promote healthy, sustainable cities.

Prediction of decay heat using non-destructive assay

This research paper introduces a novel approach to predict the decay heat of spent nuclear fuel assemblies (SNFs) using data from non-destructive gamma and neutron measurements, addressing the challenge of ensuring safety in geological repositories. Because calorimetric measurements are time-consuming, it is envisioned that gamma and neutron measurements can be used for decay heat prediction before encapsulation. This paper analyses gamma and neutron data to extract key features, specifically the activities of Cs-137, Eu-154, and the total neutron count rate. A Gaussian process model is then employed to estimate SNF decay heat. The methodology involves training a prediction model on a calibrated simulated dataset designed to mimic real experimental conditions closely. The model is then successfully used to predict the decay heat for unseen experimental data. The results highlight the potential of using gamma and neutron measurements for reliable decay heat prediction. It is shown that the magnitude of the relative deviation obtained is 2-4 %. Furthermore, the study explores the impact of removing certain input features or adjusting their uncertainty levels on the decay heat prediction model precision, in particular for the Eu-154 activity and neutron count rate. This comprehensive methodology paves the way for applying these techniques to a larger experimental scale offering a significant advancement in the safety assessment of SNFs prior to encapsulation and long-term storage.

Bayesian-Learning-Based Diffusion Least Mean Square Algorithms Over Networks.

To improve the learning performance of the conventional diffusion least mean square (DLMS) algorithms, this article proposes Bayesian-learning-based DLMS (BL-DLMS) algorithms. First, the proposed BL-DLMS algorithms are inferred from a Gaussian state-space model-based Bayesian learning perspective. By performing Bayesian inference in the given Gaussian state-space model, a variable step-size and an estimation of the uncertainty of information of interest at each node are obtained for the proposed BL-DLMS algorithms. Next, a control method at each node is designed to improve the tracking performance of the proposed BL-DLMS algorithms in the sudden change scenario. Then, a lower bound on the variable step-size of each node of the proposed BL-DLMS algorithms is derived to maintain the optimal steady-state performance in the nonstationary scenario (unknown parameter vector of interest is time-varying). Afterward, the mean stability and the transient and steady-state mean square performance of the proposed BL-DLMS algorithms are analyzed in the nonstationary scenario. In addition, two Bayesian-learning-based diffusion bias-compensated LMS algorithms are proposed to handle the noisy inputs. Finally, the superior learning performance of the proposed learning algorithms is verified by numerical simulations, and the simulated results are in good agreement with the theoretical results.

Symptom Network Analysis and Unsupervised Clustering of Oncology Patients Identifies Drivers of Symptom Burden and Patient Subgroups With Distinct Symptom Patterns.

Interindividual variability in oncology patients' symptom experiences poses significant challenges in prioritizing symptoms for targeted intervention(s). In this study, computational approaches were used to unbiasedly characterize the heterogeneity of the symptom experience of oncology patients to elucidate symptom patterns and drivers of symptom burden. Severity ratings for 32 symptoms on the Memorial Symptom Assessment Scale from 3088 oncology patients were analyzed. Gaussian Graphical Model symptom networks were constructed for the entire cohort and patient subgroups identified through unsupervised clustering of symptom co-severity patterns. Network characteristics were analyzed and compared using permutation-based statistical tests. Differences in demographic and clinical characteristics between subgroups were assessed using multinomial logistic regression. Network analysis of the entire cohort revealed three symptom clusters: constitutional, gastrointestinal-epithelial, and psychological. Lack of energy was identified as central to the network which suggests that it plays a pivotal role in patients' overall symptom experience. Unsupervised clustering of patients based on shared symptom co-severity patterns identified six patient subgroups with distinct symptom patterns and demographic and clinical characteristics. The centrality of individual symptoms across the subgroup networks differed which suggests that different symptoms need to be prioritized for treatment within each subgroup. Age, treatment status, and performance status were the strongest determinants of subgroup membership. Computational approaches that combine unbiased stratification of patients and in-depth modeling of symptom relationships can capture the heterogeneity in patients' symptom experiences. When validated, the core symptoms for each of the subgroups and the associated clinical determinants may inform precision-based symptom management.

FRGEM: Feature integration pre-training based Gaussian embedding model for Chinese word representation

Most Chinese word representation approaches mine the lexeme of Chinese words by predicting the target word or the contextual words. However, Chinese words may have multiple meanings, above ways may only capture one aspect implication. In this paper, for better modeling multiple connotations of Chinese words, a Feature integRation pre-training based Gaussian Embedding Model (FRGEM) is proposed with three stages, feature integration pre-training, relevancy Gaussian representation and similarity-based training. In the first stage, the internal sequences with three inner-character stroke, structure and pinyin features of Chinese words are generated to learn their semantic relevancy, then we combine BERT with internal sequences to extract sentence relevance and integrate the internal and sentence features to learn pre-training representations. In the second stage, relevancy Gaussian representation with pre-training embeddings is proposed to fuse inner-character features with Gaussian through estimating its mean to analyze multiple implications of Chinese words. In the third stage, a similarity-based objective is proposed to distinguish true contextual words and learn Chinese word representations. Extensive experiments on word similarity, word analogy, named entity recognition and text classification show that FRGEM is superior to majority state-of-the-art algorithms.

Optimization design of composite cap-shape pillar structure based on gaussian process of combined kernel function

Abstract To quickly realize the optimal design of the cap-shaped structure, an optimization design method of composite cap-shaped pillar structure based on combined kernel function Gaussian process was proposed. First, the shortcomings of two kernel functions representing different linear characteristics of data in model construction were analyzed, on which a new combined kernel function was constructed. Based on the finite element numerical simulation results, a database with the web thickness, web height, fiber volume, and ultimate bearing capacity of the cap-shaped structure was constructed. Then, a Gaussian process model based on the combined kernel function was established and the model parameters were trained by the maximum likelihood method. Finally, taking the ultimate bearing capacity per unit volume as the optimization objective, the artificial fish swarm algorithm was used to quickly obtain the structure design parameters under the optimal objective function, and the simulation results were compared with the initial design parameters. The results showed that the method proposed in this paper can optimize the structure performance, improve the design efficiency, and meet the engineering design requirements.

Research on multi-source information fusion tool wear monitoring based on MKW-GPR model

To enhance intelligent manufacturing capabilities and achieve the goal of online real-time high-precision monitoring of tool wear with small sample data, this paper proposes a multi-source information fusion method for turning tool wear monitoring based on the multi-kernel weighted Gaussian process regression (MKW-GPR) model. Firstly, the framework and process of multi-source information fusion for turning tool wear monitoring are established, and the experimental platform is constructed. Secondly, a full-life cutting experiment of the turning tool is conducted. During the experiment, tool wear image data and multi-sensor signal data, including cutting force and vibration signals, are collected. Then, the experimental data were processed to construct a dataset with multi-source information fusion, leading to the establishment of the MKW-GPR model. Subsequently, the experimental dataset was employed to validate the effectiveness of the turning tool wear monitoring model. The results demonstrate that, in recognizing the wear state, the MKW-GPR model exhibited superior recognition performance compared to the single kernel function Gaussian process regression model. Furthermore, in predicting the wear value, the MKW-GPR model is better than the long short-term memory (LSTM), random forest (RF), and scalable end-to-end tree boosting (XGBoost) models. During the small sample size priori point data prediction performance test, the MKW-GPR model achieved the highest accuracy, with a maximum relative error of only 2.8 % in predicting tool life during the severe wear stage. These findings substantiate the feasibility and effectiveness of the proposed turning tool wear monitoring method.

Diffusion behavior and environmental impact of odorants and TVOCs detected in a wastewater treatment plant for collaborative leachate treatment in Northwest China

Wastewater treatment plants (WWTPs) are major sources of volatile gaseous compounds, especially in mixed-source systems such as domestic wastewater and landfill leachate. This study aimed to investigate the emission behavior and environmental impact of gaseous substances, such as hydrogen sulfide (H2S), ammonia (NH3), carbon sulfide (CS2), and phosphine (PH3), at a WWTP in Northwest China. Odorants were detected in the air surrounding the grid room (XGS), biochemical treatment tank (SHC), secondary sedimentation tank (ECC), and sludge dewatering room (NTS). For comparison, the upwind boundary (O-SF) and downwind boundaries (O-XF) monitoring points were used, with odor concentrations ranging from 3.95 to 725.27 odor units. The concentration ranges of the odorant substances were 5.27–88.69, 5.61–71.96, 5.70–32.63, and 0.12–5.87 mg/m3 for H2S, NH3, CS2, and PH3, respectively. Meteorological factors such as temperature, relative humidity, and wind speed and direction substantially influence odorant emissions. The concentrations of various odorants and volatile organic compounds (VOCs) detected at the O-XF monitoring point were higher than those detected at the O-SF monitoring point, indicating that the wind intensified their diffusion toward the downwind plant boundary. The average odor intensities of odorant substances emitted from wastewater or sludge treatment equipment were 3.37, 5.09, 4.42, 2.00, and 3.82 for total VOCs, H2S, NH3, CS2, and PH3, respectively. Among them four, with downwind diffusion, only H2S presented olfactory and chronic toxicity risks based on Gaussian plume model calculations. The hazard index ranking across monitoring sites was XGS > NTS > SHC > ECC > O-XF > O-SF. These findings emphasize the urgent need for effective measures to control and mitigate gaseous pollutants emitted by collaborative WWTPS, thereby protecting environmental quality and public health.

Numerical analysis of Gaussian potential patches model depending on the substrate doping in inhomogeneous Schottky barrier diodes over a wide temperature range

The current-voltage (I-V-T) characteristics of an inhomogeneous n-type GaAs Schottky barrier diode have been investigated by numerical analysis using the modified thermionic emission (TE) current equation by Tung in the 40–320 K range at 40 K intervals. This total current (TC) equation consists of TE current and the patch current components. The patch current dominates through the low Schottky barrier height patches at low temperatures. From the I-V-T characteristics given for three different standard deviations (σ) at each substrate doping value Nd, we have determined the temperatures at which the patch current begins to dominate. The starting temperature of the patch current has decreased as the σ and Nd values decrease. It has been seen that the temperature at which the patch current component begins to dominate is about 120, 80, and 60 K for σ4, σ3, and σ2 at Nd=1.0×1014cm−3 or Nd=1.0×1015cm−3, respectively; 160, 120, and 80 K at Nd=5.0×1015cm−3; and 200, 160, and 80 K at Nd=1.0×1016cm−3, respectively. Moreover, for the substrate with high doping, it has been observed that the I-V curve of the patch current component or the TC shifts toward higher voltages than the expected position at low temperatures. Thus, from the I-V-T characteristics, it has appeared that Tung’s pinch-off model tends to be more applicable to lightly doped semiconductors. Moreover, the TC equation should be used at high temperatures because the I-V curves at high temperatures belong to the TE component, and the patch current expression without the TE component should be especially used for fit to the experimental curves at low temperatures.

A Novel Fault Feature Selection and Diagnosis Method for Rotating Machinery with SI-IR2CMSE and SSGMM-SR

Abstract In the field of fault diagnosis, machine learning is highly valued for its broad applicability and efficiency. Feature extraction and feature selection are key steps in the application of machine learning, and the performance of fault diagnosis methods relies heavily on the effective execution of these two steps. For this reason, this paper aims to enhance the performance of fault diagnosis methods by improving these two aspects. Firstly, to address the non-linearity and non-stationarity of rotating machinery vibration signals under variable operation conditions, this paper proposes an improved rapid refined composite multiscale sample entropy (IR2CMSE) feature extraction method. In addition, this paper decomposes the vibration signals with improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) and extracts the sensitive intrinsic modal functions’ (IMFs') IR2CMSE values (SI-IR2CMSE) as the initial feature vector, which more accurately reveals the intrinsic time-scale characteristics of the vibration signals. Secondly, to address the problem of over-reliance on sample labels in most feature selection methods, this paper proposes a semi-supervised Gaussian mixing model with sparse regularization (SSGMM-SR) feature selection model. The model does not require complete fault labels and can automatically identify important features. Finally, validation with two rotating machinery fault datasets shows that the method proposed in this study exhibits high diagnostic accuracy and stability across multiple classifiers.