28 results about "Bayesian melding" patented technology

The invention discloses an oil reservoir history fitting method based on graph lasso and set smooth multi-data assimilation, and relates to the technical field of oil and gas field development. The method comprises the following steps: firstly, acquiring to-be-optimized model parameters and observation data of a target oil reservoir, constructing an observation error covariance matrix and initializing a model parameter set, performing numerical simulation by utilizing an oil reservoir numerical simulator to obtain prediction data, constructing a joint state matrix, performing dimensionless standardization processing on the joint state matrix to obtain an empirical correlation coefficient matrix, and then, performing optimization on the joint state matrix. And executing a graph lasso algorithm combined with an extended Bayesian information criterion to obtain an optimal dimensionless sparse precision matrix, extracting correlation coefficient sub-blocks required by Kalman updating from the optimal dimensionless sparse precision matrix based on a Scherr's theorem, obtaining a robust data auto-covariance matrix through a reverse reduction physical quantity outline, calculating Kalman gain in combination with an observation error covariance matrix, and calculating a Kalman filter. And the model parameter set is updated until the preset condition is met, the reservoir history fitting model parameter set is output, and the stability and precision of reservoir automatic history fitting are improved.

The invention provides a construction interruption risk reasoning and plan generation method and device. The method comprises the following steps: acquiring multiple types of risk factors influencing construction interruption and a historical data set; taking a Bayesian information criterion as a scoring function, and constructing a directed acyclic graph structure based on the scoring function, a historical data set and a constraint set determined by domain knowledge; the constraint set is used for limiting whether edges between nodes exist or not in the construction process of the directed acyclic graph structure; based on the directed acyclic graph structure and the historical data set, a conditional probability table of the directed acyclic graph structure is determined, and the directed acyclic graph structure and the conditional probability table form a Bayesian network model; determining the risk probability of the target construction interruption event based on a Bayesian network model; and optimizing a preset construction plan combination based on the risk probability, and determining an optimal plan combination. According to the directed acyclic graph structure, objective data and domain knowledge are considered, and the accuracy of subsequent risk reasoning and plan generation is improved.

The invention discloses a multi-dimensional night urination behavior monitoring system based on rhythm features, and relates to the technical field of biomedical signal processing, and the system specifically comprises a data acquisition module, a rhythm feature extraction module, a urine volume dynamics modeling module, a clustering monitoring module and a real-time feedback module; collecting individual data in real time through IoT equipment and preprocessing the individual data; calculating a nocturnal urination frequency index, a nocturnal urination time concentration ratio, a urination interval rhythm variation coefficient and a deviation index based on the individual data, and performing rhythm feature extraction; performing quadratic polynomial least square fitting on the night accumulated urine volume, and calculating a urine volume acceleration index by using an obtained second derivative to quantify a urine volume generation trend; the method comprises the following steps: constructing and preprocessing a night urine multi-dimensional digital phenotypic vector matrix, fitting a Gaussian mixture model based on an expectation maximization algorithm of a Bayesian information criterion, automatically mapping individuals into four types of subtypes according to cluster center features, and outputting individual subtype labels and confidence coefficients; and obtaining a comprehensive risk score through normalized risk assessment.

The invention belongs to the technical field of load prediction, and particularly relates to an external load dynamic prediction method. Comprising the steps that S1, different driving working conditions are simulated through a virtual prototype, a dynamic load data set of the multi-crawler walking device is obtained, and the dynamic load data set is a set of dynamic load values of the multi-crawler walking device walking on the ground at each moment and characteristic parameters related to the dynamic loads; s2, performing data preprocessing on the dynamic load data set to obtain an initial structured data set; obtaining an optimal time window length based on a Bayesian information criterion, and obtaining a recurrent neural network prediction model in combination with the initial structured data set; and S3, obtaining a to-be-detected characteristic parameter sequence based on a rolling time window method, and inputting the to-be-detected characteristic parameter sequence into the recurrent neural network prediction model for processing to realize dynamic load online prediction of the multi-crawler walking device. The application range and the calculation efficiency of the dynamic load prediction method are improved to the maximum extent.

This invention relates to a method for highlighting spike discharge data of intracranial electroencephalogram (EEG) signals. The method involves collecting background noise data from the patient's brain without neuronal discharges, preprocessing the background noise data, automatically selecting the optimal order of an autoregressive (AR) model using the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC), estimating the AR model coefficients using the Yul-Walker equation, and constructing a background noise model. The intracranial EEG signals to be processed are then subjected to high-pass filtering. A short-time Fourier transform (STFT) and window-based frame-by-frame processing strategy are used to subtract the spectrum of the signal from the noise. By adjusting the parameters of the spectral subtraction, noise removal and preservation of neuronal signal features are achieved.

The invention discloses a biomolecule relation modeling method based on semantic consistency hypergraph contrast learning, and aims to model a complex relation between biomolecules by effectively combining a local structure and global semantic information. According to the method, firstly, local representation of nodes is enriched by expanding a message aggregation mechanism guided by a subgraph, and global context information is combined to improve the modeling precision of the biomolecular relationship; secondly, performing hypergraph reconstruction by adopting a double-layer consistency mechanism, and maintaining the consistency of a local structure and semantics at a node level and a hyperedge level through semantics and structure constraints; besides, soft clustering is carried out by using a Gaussian mixture model (GMM) and a Bayesian information criterion (BIC) strategy, hyperedge selection is further optimized, and structural consistency in a reconstruction process is ensured. In order to ensure the consistency of local and global semantics, the method introduces a multi-granularity comparison target in a comparison learning framework, performs training at node, hyperedge and extended subgraph levels, feeds back high-level semantic information to low-level semantic information through a cross-layer feedback mechanism, stabilizes the model and enhances semantic alignment between representations. Experimental results show that the provided framework is excellent in performance in node classification and clustering tasks on multiple reference data sets, and compared with an existing method, local and global semantic relationships can be better captured and aligned, and the hypergraph learning effect is remarkably improved.

The invention discloses an automatic division method and system for a satellite constellation orbital plane, and relates to the technical field of spaceflight measurement and control. The method comprises the following steps: firstly, calculating an orbital plane unit normal vector based on a satellite orbit inclination angle and an ascending node right ascension, and performing first-level spatial pointing clustering by adopting a density clustering algorithm to form an orbital plane group; secondly, in the same orbital plane group, by taking a semi-major axis and an eccentricity rate as characteristics, automatically determining an optimal shell layer number by utilizing a Gaussian mixture model and a Bayesian information criterion, and realizing second-stage orbital energy shell layer division; and finally, in the same shell layer, sorting and interval statistical analysis are performed on the right ascension of the ascending node, recursive subdivision is performed through dynamic threshold segmentation in combination with a range-based dual-condition verification mechanism, and fine division of a third-level orbital plane is completed. The method realizes full-process automation from orbit data to division results, has the advantages of high division precision, strong adaptability, good engineering practicability and the like, and is suitable for operation and maintenance management of large-scale heterogeneous satellite constellations.

The present application relates to the technical field of image data processing, and more particularly to a concrete vibration quality evaluation method and system based on image analysis. The method comprises: acquiring a gray image of the surface of the newly vibrated concrete; constructing a region adjacency graph and extracting multi-dimensional features; calculating physical entropy weight based on the multi-dimensional features; performing region merging operation; identifying bubble clusters and calculating the quality evaluation index of the concrete surface according to the statistical characteristics of the bubble clusters; and outputting the concrete vibration quality evaluation result according to the quality evaluation index. The present application divides the concrete surface by the watershed algorithm and region adjacency graph, extracts multi-dimensional features, constructs a physical entropy weight model to evaluate the saliency, uses a boundary gray level priority queue and an adaptive multi-threshold intelligent region merging method to solve the problem of over-segmentation, and combines the Gaussian mixture model and the Bayesian information criterion to identify bubble clusters, so as to comprehensively evaluate the concrete vibration quality and improve the objectivity and accuracy.

This application discloses a method, apparatus, equipment, and medium for assessing voltage sag fault risk, belonging to the field of voltage sag fault risk assessment. The method involves: collecting test points of the withstand characteristic curves of the electrical equipment under test under different voltage sag initiation angles and load conditions; each test point includes residual voltage and duration values; constructing multiple first Gaussian mixture models with different numbers of Gaussian components based on the test points; iteratively solving the optimization parameters of each model using the expectation-maximization algorithm until convergence, obtaining multiple second Gaussian mixture models; selecting a target Gaussian mixture model using the Bayesian information criterion; and finally, inputting the voltage sag test points to be assessed into the target Gaussian mixture model to obtain the assessment result characterizing the voltage sag fault risk of the electrical equipment under test. Therefore, by implementing this application, the problem of low accuracy in voltage sag fault risk assessment in the prior art can be solved.

This invention relates to the field of optical fiber communication testing technology, specifically a method for distinguishing adjacent fault points and estimating the distance between them using an OTDR. The method first acquires the OTDR backscattering curve of the optical fiber link under test, and then locates the starting position of the first fault point through preprocessing. Within a local analysis window, single-fault and double-fault physical models considering the system's finite pulse width and bandwidth characteristics are established. After baseline removal processing of the local observation signal, the optimal parameters of the two models are estimated using ordinary least squares and variable projection methods, respectively. Then, model selection is performed based on the Bayesian information criterion to determine the number of fault points, and when a double fault is identified, the estimated distance between the two fault points is output. This invention solves the problem of accurately distinguishing the number and estimating the distance between adjacent fault points when their responses overlap, significantly improving the OTDR's ability to distinguish and locate nearby fault points. It has high computational efficiency, good robustness, and is suitable for long-distance, large dynamic range optical fiber link testing scenarios.

The invention belongs to the technical field of electric digital data processing, and particularly relates to an electric power marketing business transaction diagnosis method based on dynamic self-adaption and feature fusion, which comprises the following steps: acquiring electric power marketing work order data, respectively extracting multi-scale features, and obtaining high-dimensional feature vectors; clustering the high-dimensional feature vectors by adopting a Gaussian mixture model, determining the number of scenes through a Bayesian information criterion, and performing dynamic evolution recognition on the scenes to obtain a scene division result; independently training an exclusive normal mode model for each divided scene; and for newly input power marketing work order data, calculating transaction scores by using the normal mode model corresponding to the scene, judging transaction, positioning features corresponding to transaction during transaction, analyzing a service root cause, and outputting a transaction diagnosis result. According to the method, accurate detection of power marketing work order transaction is realized through multi-scale feature fusion and dynamic scene adaptive modeling, and both technical accuracy and service practicability are considered.

This disclosure provides an indoor temperature clustering method and apparatus for centralized heating systems, applicable to the field of precise control of centralized heating systems. The method includes: generating initial room temperature sequence data based on room temperature data acquired from multiple room temperature sensors; constructing room temperature observation vectors of a predetermined duration based on target room temperature sequence data, wherein the target room temperature sequence data is obtained by preprocessing the initial room temperature sequence data; determining a Bayesian information criterion value based on the number of room temperature observation vectors of the predetermined duration, and determining the number of model parameters corresponding to the minimum value of the Bayesian information criterion value as the target number of clusters; performing cluster analysis on the room temperature observation vectors of the predetermined duration according to the target number of clusters, and displaying the clustering results obtained when the cluster analysis converges.

The invention discloses a power system operation state evolution analysis method, device and equipment and a medium, and the method comprises the steps: obtaining the output data of various generator sets in a plurality of geographic regions and the power load data corresponding to each geographic region, so as to generate a power system operation time sequence; outputting a state transition probability matrix through a pre-trained hidden Markov model according to the operation time sequence of the power system; wherein the hidden state number of the hidden Markov model is determined according to an akaike information criterion and a Bayesian information criterion so as to ensure the optimality of the model; the state transition probability matrix is used for identifying an implicit operation state in the operation process of the power system; and acquiring real-time operation data of a target power system, and predicting an operation state corresponding to a preset future time sequence according to the real-time operation data and the state transition probability matrix. Compared with the prior art, the dynamic property and the accuracy of the evolution analysis of the operating state of the power system can be improved.

The application provides a single-channel mechanical vibration signal blind separation method, system, device and medium, a single vibration sensor is used to obtain a mechanical vibration signal under a running condition as a single-channel mechanical vibration signal to be analyzed, the single-channel mechanical vibration signal is decomposed by using symplectic geometry modal decomposition to obtain a single-component signal, the single-component signal is reconstructed according to a frequency spectrum correlation coefficient to obtain a symplectic geometry modal component, the single-channel mechanical vibration signal, the symplectic geometry modal component and a residual term are used to form a multi-dimensional observation signal, a singular value decomposition is performed on a self-correlation matrix of the multi-dimensional observation signal, a Bayesian information criterion is used to obtain an estimation of the number of signal sources, the symplectic geometry modal component is selected according to a maximum time-domain correlation coefficient principle to form a new multi-dimensional signal observation matrix together with the original observation signal, and a blind source separation method based on time-frequency analysis is used to separate the multi-dimensional signal observation matrix. The application can separate multiple mechanical vibration source signals from a single-channel vibration signal.

The source number determination method, device, equipment and storage medium provided by the present disclosure determine a data matrix according to a received signal, calculate a sample covariance matrix of the data matrix and perform eigenvalue decomposition, calculate a linear spectrum statistic, calculate a generalized Bayesian information quantity corresponding to each source number, sort the generalized Bayesian information quantities from large to small, and determine the source number corresponding to the smallest generalized Bayesian information quantity as the estimation of the source number. By converting the original data matrix into a linear spectrum statistic, the main information contained in the data matrix is extracted, the description framework of the asymptotic behavior of the statistic is changed, and the requirement of accuracy for the signal-to-noise ratio is reduced through the dimension ratio p / n.

The invention provides a parameter estimation method and system for a cross-field ultrahigh-frequency wireless channel, and belongs to the field of wireless communication, and the method comprises the steps: constructing an ultrahigh-frequency terahertz channel measurement system, and obtaining a channel transmission function; establishing a cross-field domain channel signal model; executing a cross-field domain parameter estimation and optimization process by adopting a parameter estimation algorithm based on a maximum likelihood criterion; and carrying out channel characterization and analysis based on the estimation parameters. According to the method, near-field and far-field multipath components are intelligently distinguished by introducing the Bayesian information criterion, parameter estimation is performed by adopting the spherical wavefront model and the plane wavefront model respectively, and the problem of estimation deviation caused by neglecting a near-field effect and spatial non-stationarity in a traditional algorithm is effectively solved by combining iterative optimization and visible region modeling. According to the method, the nonlinear phase change, the multipath birth and death phenomenon and the frequency domain characteristic in the ultrahigh-frequency multi-antenna channel can be accurately captured, the parameter estimation precision and the residual performance are remarkably improved, and a key technical support and an experimental basis are provided for channel modeling, large-scale antenna array design and perception communication integration of a 6G ultrahigh-frequency terahertz communication system.

The application discloses a tower control visual field efficiency optimization method based on improved gaze point rendering, relates to the technical field of air traffic control simulation, and comprises the following steps: acquiring objects in a current scene, control task data, interactive data of a controller and a simulated tower system, and a current gaze point rendering picture; performing feature extraction on the objects in the current scene and the control task data; inputting the extracted feature data and the interactive data into a preset Bayesian information gain framework to output an attention distribution of a current controller; calculating parameters of a Gabor kernel based on the Gabor kernel function according to the attention distribution of the current controller; generating Gabor noise based on the parameters of the Gabor kernel, superimposing the Gabor noise on the current gaze point rendering picture, and forming a current final rendering image. The application significantly improves the rendering efficiency and enhances the immersion and visual reality of a user.

The application discloses a micro-architecture design space exploration method based on Gaussian mixture regression, adopts a Bayesian optimizer to realize incomplete supervised learning, can reduce the labeling cost of a data set, and accelerates the training of a model by utilizing prior probability, adopts a Gaussian mixture regression model as a proxy model, can simultaneously calculate the predicted mean of a target value and the covariance between multiple targets, compared with other models, can better perform multiple target optimization, uses a Bayesian information optimization criterion to determine the final number of Gaussian components, realizes the function of dynamically optimizing the number of Gaussian components, and makes the model better approximate a real function, collects a function CEIPV in combination with the target value, considers the balance problem between targets while providing the observation points with the highest uncertainty, and realizes the operation of selecting an optimal micro-architecture outside the data set by a conjugate gradient method.

The application discloses a hydrological time series cycle identification method based on spectral peak guidance and iterative waveform matching, which comprises the following steps: obtaining a stationary sequence based on the Akaike information criterion adaptive detrending; iteratively extracting cycles, using residual power spectrum in each round, retaining strong spectral peaks and generating real candidate cycles in the neighborhood to reduce the candidate set capacity; adopting phase binning median interpolation to construct a non-parametric waveform, correcting significance test by multiple comparison, and screening the optimal cycle; correcting the amplitude by joint least squares fitting, determining acceptance according to the Bayesian information criterion drop, and determining continuous extraction by white noise test; and finally outputting the significant cycle and the corrected component. Through the iterative process of spectral peak guidance, waveform matching, residual updating and re-guidance, the application can automatically identify real cycles and complex waveforms, suppress long cycle artifacts and over-extraction, and improve the accuracy and robustness of hydrological cycle identification.

A monocrystalline silicon grinding contact state monitoring method based on Bayesian information amount and modal decomposition comprises the steps that acoustic emission AE signals in the grinding process are collected, and multi-scale decomposition is conducted through discrete wavelet decomposition (DWT), local mean decomposition (LMD) and empirical mode decomposition (EMD); constructing a double-section constant Gaussian model for each component signal, calculating a Bayesian information criterion BIC value of each candidate segmentation point, and positioning a contact moment through a Bayesian information criterion BIC curve slope minimum value; finally, an IMF (intrinsic mode function) result most sensitive to mutation in EMD (empirical mode decomposition) is optimized to serve as judgment output; compared with the prior art, the method does not need to preset a threshold value, is high in noise immunity, can capture the contact starting point at sub-millisecond precision, effectively overcomes the limitation of a single decomposition method, and remarkably improves the detection accuracy, robustness and engineering applicability.

The invention provides a power transaction market normalized operation supervision method and system and a storage medium, and the method comprises the steps: calculating a market deviation index as an endogenous variable of a vector autoregression model through obtaining the time series data of a unit quotation, a market clearing price, a power transmission section power flow, a renewable energy power generation amount and a system total load; a Bayesian information criterion is adopted to screen an initial lagging order, a final lagging order is determined in combination with an operation period extracted by Fourier transform, a two-dimensional state space is constructed based on a renewable energy power generation proportion and a market deviation index, market operation state intervals are divided through unsupervised clustering, and a new energy power generation mode is established. And a model coefficient and an error covariance matrix in each state are independently solved, a corresponding model is selected according to the current state to carry out multi-step prediction, a multi-dimensional prediction confidence ellipsoid reflecting prediction uncertainty is generated, and when a market actual operation state vector falls outside the confidence ellipsoid, a supervision signal is generated to identify that the market deviates from a normal state.

This invention relates to a method for predicting financing demand based on a heterogeneous clustering distributed lag model. The method includes the following steps: constructing a set of predictive and response variables for investors; establishing a heterogeneous clustering distributed lag model and training it using an improved K-means algorithm, determining the number of clusters using the Bayesian information criterion; and performing out-of-sample prediction based on the trained model to predict changes in the financing balance under different financing interest rate adjustment schemes. The method, apparatus, processor, and computer-readable storage medium for predicting financing demand based on a heterogeneous clustering distributed lag model of this invention aim to utilize economic and econometric models, combined with customer credit information and macroeconomic indicators, to predict the likelihood of customers adjusting financing interest rates and the changes in the financing balance after interest rate reductions. By identifying the customer group truly affected by interest rate changes, this helps companies adjust interest rate strategies more accurately and improve the efficiency of financing balance management.

The invention relates to the technical field of image data processing, in particular to a concrete vibration quality evaluation method and system based on image analysis. The method comprises the following steps: acquiring a grayscale image of a newly vibrated concrete surface; constructing a region adjacent graph and extracting multi-dimensional features; calculating a physical entropy weight based on the multi-dimensional features; carrying out region merging operation; identifying the bubble clusters, and calculating quality evaluation indexes of the concrete surface according to statistical characteristics of the bubble clusters; and outputting a concrete vibration quality evaluation result according to the quality evaluation index. According to the method, a concrete surface is segmented through a watershed algorithm and a region adjacent graph, multi-dimensional features are extracted to construct a physical entropy weight model to evaluate significance, a boundary gray priority queue and a self-adaptive multi-threshold intelligent merging region are utilized, excessive segmentation is solved, bubble clusters are identified in combination with a Gaussian mixture model and a Bayesian information criterion, and the accuracy of bubble cluster recognition is improved. Therefore, the concrete vibration quality is comprehensively evaluated, and objectivity and accuracy are improved.

The application relates to a user clustering method, device, equipment, storage medium and program product. The method comprises the following steps: obtaining a target Gaussian mixture model by adjusting probability distribution functions contained in an initial Gaussian mixture model according to sub-probability values of historical load curves of a user to be clustered in a historical period and belonging to each probability distribution function in the initial Gaussian mixture model; determining a target load curve corresponding to the user to be clustered according to sub-probability values of the historical load curve corresponding to the user to be clustered and belonging to each probability distribution function in the target Gaussian mixture model; clustering the user to be clustered according to the target load curve corresponding to the user to be clustered; and judging whether each user after clustering reaches a clustering end condition according to Bayesian information criterion values corresponding to each user after clustering; and if yes, obtaining a clustering result of the user to be clustered. The method can improve user clustering accuracy.

The application provides a kind of grain security resilience influence factor analysis method based on Bayesian network.The method relates to the field of grain security.The method comprises the following steps: by Spearman rank correlation coefficient method, adjusted R 2 The maximum criterion optimal subset, random forest feature importance evaluation three kinds of methods are used to screen the influence factors of grain security resilience, and the union method is used to determine the index; the natural break point method is used to convert the quantitative index into three types of qualitative index by minimizing the within-group variance and maximizing the between-group variance; the grain security resilience is used as the dependent variable, and the final index is used as the independent variable; the simulated annealing algorithm is used to construct a three-layer Bayesian network; the Bayesian information criterion scoring function is used to evaluate the iteration to obtain the best structure; based on this, the reverse reasoning method is used to calculate the reverse reasoning probability and change of each index, and the influence effect is analyzed. The application can systematically and comprehensively analyze the influence effect of various factors.

The present application relates to the technical field of power system, specifically relates to a power system low-frequency oscillation identification method based on UBSS algorithm, steps are as follows: through space-time conversion, real-time observation information mining is realized, and the observation multi-channel is expanded, and the UBSS problem is converted into BSS problem;The source number estimation of observation information is carried out using the Bayesian information criterion, the number of dominant oscillation modes is determined, the modal order is realized, and the modal decomposition task is completed using the BSS algorithm, and the estimation of oscillation mode parameters is realized.This method can decompose the main component in the complex signal under the condition of underdetermination, compared with HHT and Prony, has better noise robustness;In the calculation error of parameter estimation accuracy, the performance of the method is better, and dynamic information of modal parameters can be provided;Based on the observation signal, the number of low-frequency oscillation modes can be accurately estimated in the absence of system topology, the dependence on system information is reduced, and the applicability of the method is improved.

The application discloses a low-frequency non-intrusive load monitoring method and system based on adaptive event detection, and the method comprises the following steps: event detection, feature extraction and load identification; in the event detection, the Bayesian information criterion is used as a detection window, and the window threshold is adaptively optimized through a power variable point weight model; the power time sequence of different electrical equipment is decomposed and grouped for load features by using a variational mode decomposition method; and according to a load identification model, the load identification and classification of the power curve waveform diagram of different electrical equipment are realized. The method disclosed by the application guarantees high precision of event detection and also shows high load identification accuracy under the condition of low-frequency sampling.

The application discloses a kind of repairable equipment reliability modeling and evaluation method considering incomplete repair.By using three-parameter boundary intensity process model (3-BIP) as benchmark failure intensity function to correct proportional hazard intensity (LPIM) model, the reliability of repairable equipment is modeled;The model parameter estimation problem is converted into the maximum value solving problem of nonlinear objective function, and the particle swarm optimization algorithm is used to solve the model parameters;Thirdly, the inverse Fisher information matrix method is used to estimate the confidence interval of model parameters, and the point estimation and interval estimation method based on Delta method of key reliability indicators are given;The application is applied to the fault truncated time data of a wind generator, and the Akaike information criterion, Bayesian information criterion and goodness-of-fit index are used to test the goodness of fit of the model.The goodness of fit of the model is better than BIP and LPIM model, and the reliability evaluation result obtained is more in line with engineering practice.