30 results about "Statistical correlation" patented technology

This invention relates to a geochemical prospecting method based on graph neural networks and spatial features, enabling geochemical prospecting under different geological environments based on accurately predicted background concentrations of different chemical elements. The method includes: preprocessing raw geochemical multi-element data and separating it from background data to obtain background area data; constructing spatial context statistical features for each sampling point of the geochemical multi-element data according to a preset spatial scale, and fusing them with the original environmental features of that sampling point to generate an enhanced environmental feature vector for each sampling point; constructing a graph topology based on the statistical correlation between chemical elements in the background area data using the maximum spanning tree algorithm; constructing a graph neural network based on the graph topology, inputting the enhanced environmental feature vector of each sampling point into the graph neural network, outputting the predicted background concentration of each chemical element at each sampling point, and calculating a mineral prediction map.

The application discloses a piezoelectric guided wave online damage alarm method for an aircraft test flight structure, and belongs to the field of aircraft structure health monitoring, and comprises the following steps: constructing a guided wave sensing network; obtaining a reference feature sample set by acquiring a reference feature through a variable step window sliding method; performing probability statistical modeling on the reference feature sample set, obtaining a probability statistical correlation feature and a reference feature sample set residual error, and taking an absolute value to obtain a reference alarm feature; establishing a monitoring feature sample set, multiplying the monitoring feature sample set and the probability statistical correlation feature to obtain a monitoring feature sample set residual error, and taking an absolute value to obtain a structure damage alarm feature; performing network weighted fusion on the structure damage alarm feature, determining a damage high-probability channel according to a channel screening threshold, calculating a structure damage alarm threshold of the damage high-probability channel, and realizing structure damage alarm. The method is simple and efficient in implementation process, and realizes reliable alarm of structure pit damage under the influence of time-varying factors.

This invention provides a method, apparatus, device, and storage medium for inferring a single-sample ncRNA regulatory network, relating to the field of gene identification technology. The specific implementation includes: acquiring ncRNA and target gene transcriptome data of a matched sample; for each target sample in the matched sample, calculating a first regulatory relationship strength matrix and a second regulatory relationship strength matrix of ncRNA and target gene before and after removing the target sample using a preset statistical algorithm; obtaining the fusion regulatory relationship strength matrix of ncRNA and target gene corresponding to the target sample based on the first and second regulatory relationship strength matrices; and inferring the single-sample ncRNA regulatory network of the target sample based on the fusion regulatory relationship strength matrix of ncRNA and target gene corresponding to the target sample. This invention can reflect the regulatory relationship strength between ncRNA and target gene at the single-sample level by constructing statistical correlation values, thereby inferring the single-sample ncRNA regulatory network.

The application discloses a transaction strategy simulation evaluation system based on scenario generation, and belongs to the technical field of financial transactions. The system comprises a parameter setting module, a data acquisition module, an economic scenario generation module and a strategy evaluation module. The economic scenario generation module constructs macro, meso and micro models through hierarchical modeling, couples each model by using a joint correlation matrix, and generates multiple future market scenario paths based on Monte Carlo simulation. The strategy evaluation module executes the strategy on each path and calculates the evaluation index. The system realizes joint dynamic simulation of multi-dimensional economic variables and asset yield, solves the technical problem that traditional models are difficult to depict cross-level statistical correlation, and improves the consistency and reliability of the simulation path. Through end-to-end automatic integration, manual intervention and system interaction overhead are significantly reduced, efficient strategy simulation testing of large scale and multiple scenarios is supported, and system computing throughput and testing efficiency are improved.

The application discloses a short-term strong earthquake occurrence time prediction method and device, medium and equipment, and relates to the technical field of earthquake prediction. The method performs standardization preprocessing on atmospheric tide gravity signals in a preset time window, inputs a plurality of target days into a prediction model constructed in advance after detecting an atmospheric tide gravity anomaly event, and obtains the probability of strong earthquakes occurring in each target day after the atmospheric tide gravity anomaly event. The prediction model is trained according to the statistical correlation characteristics between the atmospheric tide gravity anomaly event occurrence time and the strong earthquake occurrence time. The strong earthquake refers to an earthquake with a magnitude greater than a preset threshold. According to the probability of strong earthquakes occurring in each target day after the anomaly, the strong earthquake occurrence time prediction result is determined. The method realizes quantitative probability prediction of the strong earthquake occurrence time, and has the advantages of being interpretable, calibratable and online updatable.

This invention relates to the field of landslide dammed lake emergency response technology, and discloses a method and application for identifying landslide dammed lake-prone areas by integrating fuzzy frequency ratio and random forest methods. The method includes: dividing the target area into evaluation units; constructing an evaluation index system for landslide dammed lake-prone areas of the target area based on a random forest model; calculating the landslide sensitivity weights of the evaluation indicators based on the fuzzy logic frequency ratio analysis method; calculating the river blockage probability weights of the evaluation indicators based on a judgment matrix; fusing the landslide sensitivity weights and river blockage probability weights of the evaluation indicators to determine the combined weights of the evaluation indicators; and extracting landslide dammed lake-prone area information based on weighted superposition analysis. This invention employs a fuzzy logic-based frequency ratio analysis method, analytic hierarchy process (AHP), and weighted superposition analysis method to obtain indicator weights, fully integrating the statistical correlation between evaluation indicators, expert experience, and decision-makers' personal preferences, thereby improving the accuracy of the landslide dammed lake-prone area evaluation model.

This invention relates to the field of electricity theft identification technology in the power industry, specifically a method for identifying electricity theft users on 10 kV non-economically operating lines. The method first collects daily line loss rate and daily electricity consumption data from users, performing preprocessing including missing value repair and outlier removal using box plots. Next, it calculates the Pearson correlation coefficient between the line loss rate and user electricity consumption, constructs an adaptive anomaly judgment threshold range based on the statistical characteristics of the correlation coefficient, verifies the significance of the correlation coefficient through hypothesis testing, and finally combines the threshold range with the test results to perform a joint judgment to identify suspected electricity theft users. This invention identifies the hidden correlation between line loss rate and user electricity consumption from a statistical correlation perspective, overcoming the limitations of traditional methods that rely on sudden changes in line loss rate. It can accurately capture behaviors such as diversion and slow electricity theft, achieving high detection accuracy and strong engineering practicality, providing technical support for electricity theft prevention work.

This application discloses a coal facies logging classification method, system, equipment, storage medium, and program product, belonging to the field of coalfield geological exploration technology. First, by selecting reference wells and collecting multi-source data, a reliable data foundation is provided for coal facies classification, avoiding the limitations of subjective assignment. Second, by evaluating the correlation between coal facies parameters and logging data, logging data categories that conform to geological laws are selected, effectively eliminating noisy data irrelevant to coal facies types and improving the effectiveness of the training dataset used for model training. By determining the significance between logging data and coal facies types, a reliable statistical correlation between the logging data used for model training and coal facies types is ensured, improving the quality of model training and the accuracy of prediction. Finally, by training a coal facies evaluation model and applying it to prediction wells, the objectivity and accuracy of coal facies classification are guaranteed. This solves the technical problems of strong subjectivity and low classification efficiency in existing technologies.

The application discloses a fuzzy multi-label inference learning method for heterogeneous data, comprising: label-aware enhanced feature construction, selecting a neighbor based on a hybrid difference metric and constructing a label-aware enhanced feature vector according to a difference relationship of positive and negative neighbor weight cumulative values; fusion enhanced antecedent construction, splicing the original feature and the enhanced feature and then obtaining a fuzzy feature vector through fuzzy rule activation; label structure guided consequent alignment, constructing a label difference matrix based on a statistical correlation coefficient between labels and taking the label difference matrix as a structured alignment constraint term acting on a consequent parameter matrix; joint optimization and prediction, integrating a data fitting term and the structured alignment constraint term into a unified objective function for optimization and solution to generate a multi-label prediction result. The application solves the problems of feature mapping modeling difficulty, insufficient learning of scarce labels and label dependency relationship interference in the heterogeneous data scene, and improves the accuracy and stability of multi-label classification.

This invention discloses a virtual power plant system scheduling method based on deep reinforcement learning, belonging to the field of power energy optimization management technology. The method includes: collecting multi-source real-time data from the virtual power plant; constructing a causal graph model and generating causal feature vectors and adjacency matrices; inputting the causal feature vectors and adjacency matrices into a reinforcement learning framework, performing enhancement processing through a causal attention mechanism to generate an optimized scheduling strategy; extracting economic and safety objective function values ​​from the optimized scheduling strategy, constructing a multi-objective optimization problem, and using gradient descent to track equilibrium points and obtain a flexible scheduling scheme; inputting the flexible scheduling scheme into a high-fidelity digital twin model, performing behavioral risk detection during simulation operation, and outputting a risk report and correction parameter vectors. This invention achieves a leap from statistical correlation analysis to causal reasoning, solving the problem of blind spots in scheduling decisions under high-dimensional uncertainty.

The application discloses a power adjustable potential prediction and decision support system fusing load identification and a method thereof, and belongs to the technical field of power energy management, which comprises a data acquisition module for acquiring relevant data, wherein the relevant data comprises power load of users in a target area, relevant statistics based on the power load, and a database generated by the relevant statistics, wherein the relevant statistics comprises 10 users with the maximum power load in the target area; a data processing module for processing the database, wherein the relevant processing comprises acquiring time periods corresponding to the maximum power load of the 10 users, and acquiring three time periods with the most occurrences in the time periods; and the like. The application can effectively improve the stability and reliability of the power system, and promote the efficient use of renewable energy by acquiring and analyzing the power load of the users and the power generation data of the renewable energy, dynamically setting the load normal interval, and quantifying the influence of the fluctuation of the power generation power on the load.

The application discloses a kind of fusion SAW temperature and UHF partial discharge fault early warning method and system, belong to power equipment state monitoring technical field, it includes obtaining power equipment acoustic surface wave temperature signal and ultra-high frequency partial discharge signal, generates original time series data, generates temperature change rate parameter and partial discharge activity intensity parameter respectively, and calculates state correlation coefficient;State correlation coefficient is compared with linkage threshold value, and generates hierarchical fusion early warning signal;Analysis hierarchical fusion early warning signal, generates and outputs equipment fault early warning information containing fault nature inference.The way that temperature and partial discharge signal are synchronously collected, the statistical correlation of the time series of two is calculated and analyzed, and linkage threshold value is constructed to generate hierarchical fusion early warning signal, can accurately identify thermal-electric coupling type fault, and infer the nature of fault, improve the reliability and intelligent level of early warning.

This invention discloses a medical image classification method and device based on Brownian distance covariance and statistical dependence. It designs a statistical dependence fusion DMF module, which enhances the overall distribution pattern of perceived features through covariance statistics, quantifies the statistical correlation between local and global features through a mutual information estimator, and focuses on lesion edges and complex texture areas through standard deviation spatial attention, achieving simultaneous spatial multi-scale fusion and statistical distribution perception. A dual-branch dynamic residual fusion framework is designed, introducing a Brownian distance covariance (BDC) branch to correct the prediction results. A dynamic weighting mechanism is designed based on prediction entropy, and a multi-level BDC progressive fusion strategy is designed, embedding BDC modules in multiple feature layers of the network. A learnable weighted fusion mechanism adaptively integrates multi-scale statistical information from local texture to global semantics, constructing a complete statistical dependence pyramid. Finally, the fused features are input to the decoder for decoding, generating classification results and achieving medical image classification.

This invention discloses a method and system for tracing the source of harmonics in a distribution network, belonging to the field of distribution network monitoring technology. The method includes: S1, constructing an initial distribution network harmonic causal graph; S2, sensing real-time changes in the distribution network topology and dynamically updating the initial harmonic causal graph to obtain an updated harmonic causal graph; S3, decomposing the updated harmonic causal graph into multiple single-frequency causal subgraphs, performing causal intervention on each subgraph to separate background harmonics from user-side harmonics, calculating the contribution of each user-side harmonic source, and outputting the tracing results. This invention uses a causal graph neural network to replace traditional statistical correlation analysis, thereby effectively distinguishing between causal relationships and spurious correlations in harmonic propagation. By decomposing the harmonic causal graph into multiple independent single-frequency causal subgraphs, accurate decoupling of broadband coupled harmonics is achieved, avoiding error accumulation caused by frequency domain decomposition.

This invention provides a method and related apparatus for assessing the implicit inference risk of sensitive power grid data, belonging to the field of power system information security and data privacy protection technology. It includes the following steps: obtaining the statistical correlation coefficients corresponding to the auxiliary features of the power grid to be assessed; using the obtained statistical correlation coefficients as input to a correlation coefficient threshold discrimination boundary model; and assessing whether the auxiliary features of the power grid constitute an implicit inference risk through the correlation coefficient threshold discrimination boundary model. Specifically, the correlation coefficient threshold discrimination boundary model is constructed using the parameters of a deep neural network inference model, including a feature weight vector and a bias scalar. This invention constructs the correlation coefficient threshold discrimination boundary using learnable parameters of a deep neural network inference model, replacing the traditional method of relying on expert experience to set a fixed threshold. This achieves an objective and adaptive assessment of the implicit inference risk of sensitive power grid data, significantly improving the accuracy and scenario adaptability of the assessment.

This invention discloses an adaptive localization method based on an ensemble paleoclimate data assimilation framework. It constructs an observation density field and updates the localization radius of model grid points. If a proxy record is located within its localization radius, its weight is calculated using a Gaspari-Cohn fifth-order polynomial. If the proxy record is located outside its localization radius, the correlation information between the climate model grid points and the proxy record is calculated. If the correlation information meets a correlation threshold, the weight of the proxy record is calculated using the correlation information; otherwise, the weight of the proxy record outside the localization radius is calculated using a Gaspari-Cohn fifth-order polynomial. Combining the statistical correlation information within the reconstruction period, a final mixed weight matrix is ​​obtained, which is then used to adjust the covariance matrix. While ensuring that each model grid point retains a certain amount of observational influence, statistical correlation is used to reduce spurious teleconnections and improve reconstruction results.

This invention discloses a construction site safety risk identification method based on multimodal spatiotemporal alignment and causal hypergraph, comprising the following steps: feature extraction from real-time collected multimodal construction site data; spatiotemporal alignment and normalization of the multimodal features to obtain a normalized multimodal state vector, and constructing a causal hypergraph based on this vector; risk evolution inference under multi-causal synergy based on the causal hypergraph to obtain a comprehensive risk score; and accurate identification of the construction site safety risk status by comparing the comprehensive risk score with a preset risk threshold. This invention solves the problem of existing construction site safety risk identification methods based on deep learning models, which rely on statistical correlation between data for risk determination, only outputting risk identification results but failing to trace and explain the root causes of the risks, leading to a significant reduction in the credibility and interpretability of the risk identification results.

This invention relates to the field of marine aquaculture technology, specifically providing a method, system, equipment, and medium for environmental monitoring in marine aquaculture. The method includes: acquiring multi-source sensor data; constructing a scene graph based on the multi-source sensor data; the scene graph using sensor monitoring indicators as feature nodes and physical associations, statistical correlations, or time-series dependencies between nodes as edges; constructing a strategy graph; the strategy graph using preset processing strategies as strategy nodes and logical relationships between strategies as edges, and attaching condition labels to each strategy node; calculating the matching degree between the scene graph and the strategy graph, and outputting environmental monitoring results or triggering corresponding processing strategies based on the matching degree. This invention significantly improves the accuracy and automation level of marine aquaculture environmental monitoring through a dual-graph structure design of scene graphs and strategy graphs and a cross-graph matching mechanism.

This invention relates to a method and system for protecting the privacy of large model inference data in a confidential computing environment. The method includes: real-time acquisition of the raw token stream generated by model inference, recording the conditional probability and generation time, and performing semantic correlation analysis to generate a sequence of raw token tuples; calculating information entropy weights based on conditional probabilities and appending them to the corresponding token tuples to generate a complete token tuple sequence; dynamically aggregating and planning the tokens based on semantic equivalence class identifiers and information entropy weights to generate a token block partitioning scheme; re-encapsulating the tokens based on the partitioning scheme to generate a token block sequence; and performing length standardization processing on each token block to generate network transmission packets with uniform length characteristics and sending them to the user terminal. This invention solves the problem that existing trusted execution environments cannot prevent network side-channel attacks by eliminating the statistical correlation of data packet timing and size through semantically preserved traffic shaping, while ensuring the semantic coherence of the output.

This invention belongs to the field of rotating machinery fault diagnosis technology, specifically disclosing a cross-condition bearing fault diagnosis method and system based on a causal hierarchical network. The method includes: acquiring bearing vibration signals and preprocessing them; for the preprocessed signals, Granger causality analysis and structural causality model learning are applied respectively to obtain Granger causality weight matrices and structural causality model weight matrices; based on the Granger causality weight matrices and structural causality model weight matrices, a causal sensing gating signal is generated using a gating mechanism; the causal sensing gating signal is used to enhance the features of the preprocessed signals; the feature-enhanced signal is input into a pre-trained causal hierarchical diagnosis network to obtain the fault diagnosis result. This invention integrates causal inference and deep learning, and effectively solves the problem of insufficient generalization ability caused by the reliance on statistical correlation in traditional methods by adaptively learning the causal representation of faults through a causal hierarchical network.

The application discloses a kind of standardization processing and abnormal tracing method and system of electric power acquisition data, it is related to electric power system technical field.First, the multi-source heterogeneous operation data of electric power equipment collected is preprocessed, then feature vector is extracted, and classification algorithm is used to identify abnormal data.Combining corresponding time series record and associated event log, correlation analysis is carried out to determine potential root clues.Based on potential root clues, time series morphological features of equipment operation parameters are extracted, and clustering algorithm is used for grouping to obtain classification abnormal mode group.Through calculating the statistical correlation strength between equipment abnormal mode in classification abnormal mode group and multidimensional historical operation data, root trigger factor is located, and abnormal tracing path graph is constructed according to the same, and main problem link is marked.Finally, decision output scheme is generated based on main problem link.The application improves the stability of electric power equipment operation and the overall reliability of electric power system.

The invention discloses an atmospheric multi-pollutant forecast correction method and system fusing multi-source observation and mode output, and relates to the field of atmospheric pollution intelligent monitoring, and the method comprises the steps: mapping ground / satellite and other multi-source observation and mode output to a unified space-time reference, and forming comparable lattice point data; performing data assimilation on the mode initial field under observation constraint to obtain an assimilation analysis field with both spatial continuity and observation consistency as a background reference; combining historical error statistics of each data source and real-time deviation information of a relative background field, adaptively generating a dynamic reliability weight according to grid points and pollutants, realizing weighted fusion of multi-source observation and the background field, and obtaining a multi-pollutant preliminary fusion field; and introducing a multi-pollutant coupling constraint interaction correction model, cooperatively correcting multi-pollutant forecast and keeping physical and chemical consistency and statistical correlation. Therefore, service-oriented multi-source consistent fusion and multi-pollutant collaborative correction are realized.

The application belongs to the technical field of data processing and intelligent analysis, and particularly relates to a substation equipment fault feature identification method based on multispectral analysis, which comprises the following steps: acquiring infrared temperature distribution data and ultraviolet photon distribution data and aligning to obtain an infrared temperature distribution map and an ultraviolet photon distribution map; calculating infrared normalized rank values and ultraviolet normalized rank values based on the cumulative distribution characteristics of coordinate point intensity information, and calculating statistical correlation strength according to local joint distribution; obtaining fault saliency according to the spatial direction consistency of gradient vectors and the tail high value distribution characteristics of the infrared normalized rank values and the ultraviolet normalized rank values, and combining the statistical correlation strength; obtaining a fault region through adaptive threshold segmentation and spatial connectivity analysis, and outputting a position and a fault saliency mean value. The application reduces the risk of weak fault features being submerged by background noise in a complex environment, and improves the accuracy of fault identification.

A statistical method and display method for a target object, and a computer device, an aircraft, a control terminal and a storage medium. The statistical method for a target object comprises: selecting a target region in a first image, wherein the target region has a target object; controlling an imaging apparatus carried on an aircraft to photograph the target region so as to acquire a second image; identifying the target object in the second image; and on the basis of the identified target object, determining statistically relevant information of the target object in the second image, wherein the resolution corresponding to the second image is greater than the resolution corresponding to the first image.

This invention discloses a deep learning-based method for predicting the disease evolution of hepatobiliary patients, belonging to the fields of medical imaging and deep learning technology. The invention first acquires multi-temporal abdominal CT images, time-series clinical laboratory indicators, and medical history follow-up data of hepatobiliary patients; it then uses a variant U-Net network with Lagrange pseudo-ordering structure constraints to accurately segment the CT images and extract radiomics features; based on the segmentation results, a three-dimensional model is constructed, and numerical simulations of hepatobiliary blood perfusion and bile fluid dynamics are performed using local entropy production theory to obtain dynamic biomechanical features; cross-dimensional feature alignment and fusion are achieved by fusing attention mechanisms and a multimodal network for complex community discovery, and then chaotic theory is introduced for phase space reconstruction to enhance dynamic features; this invention achieves an upgrade from statistical correlation to physiological causality in prediction, improving model generalization and clinical interpretability, and is suitable for prognostic assessment and intelligent early warning of disease progression in hepatobiliary diseases.

The application relates to an agent model illusion suppression method and device, computer equipment and a readable storage medium. The method comprises the following steps: constructing a fact anchor point sample set; inputting a fact input and an anti-fact input into a large language model, and extracting activation values of each layer, each time and each neuron; determining an activation difference value of a three-dimensional causal unit between the fact input and the anti-fact input and a statistical correlation value between the activation value of the three-dimensional causal unit and an illusion label, and screening out a candidate causal unit set; determining an average causal effect score of the candidate causal unit, and corresponding intervention task utility scores and semantic drift degrees; screening out a key intervention unit set from the candidate causal unit set; constructing a context state vector and defining a gating function, and performing selective suppression on the three-dimensional causal units in the key intervention unit set. Therefore, the illusion rate is effectively reduced, and the task completion ability and semantic generation stability of the model are maximally reserved.

This invention discloses a method and system for processing and analyzing asymmetric distributed evaluation data, relating to the fields of data processing and artificial intelligence. The processing and analysis method includes the following steps: data acquisition, preprocessing and de-identification, multi-level data classification, and parallel artificial intelligence modeling and analysis of the S2-processed data; in the first processing flow, structured rating data is analyzed, and rating difference calculation and multi-level clustering analysis are performed sequentially; in the second processing flow, unstructured text data is analyzed, and unsupervised topic extraction based on sentence vectors and weakly supervised sentiment classification based on pre-trained models are performed, ultimately deriving text insights through statistical correlation between topic and sentiment results; an interactive data analysis interface analyzes the text based on a pre-trained language model, generating a final analysis summary and improvement suggestions; this invention solves the problem of low signal-to-noise ratio in large amounts of asymmetric distributed data, laying the foundation for refined analysis.