20 results about "Tensor factorization" patented technology

A method includes a processing computer querying a data store for plurality of network data since a last epoch of data. The processing computer can generate matrices based on the plurality of network data and then perform tensor factorization on the matrices, to obtain latent values in the network data. The processing computer can then determine that the latent values satisfy a predetermined criterion. The processing computer can input the latent values into a model associated with the signal, wherein the model generates output data. The processing computer can then generate an alert comprising the output data and then transmit the alert to a remote server computer.

This invention presents a method for establishing large deformation viscoelastic constitutive relations based on Prony technology. It addresses the problems of existing viscoelastic constitutive models lacking universality, and the complexity and programming difficulties associated with their mathematical relationships. The invention includes: decomposing the stress tensor into the sum of the deviatoric stress tensor and the mean stress; decomposing the strain tensor into the sum of the deviatoric strain tensor and the mean strain; and providing a convolutional expression for the deviatoric stress tensor and the mean stress. The increment of the deviatoric stress tensor is expressed as a function of the deviatoric strain increment, and the increment of the mean stress is expressed as a function of the mean strain increment. The tensor form of the incremental constitutive relation is written in matrix form. This method has been programmed, and the numerical results obtained from the program agree well with the experimental results in specimen-level experiments on solid rocket motor propellant grains.

The invention relates to the field of data processing, in particular to a large-model tensor database construction method and system.The method comprises the steps that tensors are captured and clustered according to semantics, sub-modes in the clusters are mined through a Gaussian mixture model (GMM), and a shared base is constructed; during online processing, a new tensor is decomposed into a composite representation formed by a mode probability and a core tensor through multi-base projection. According to the representation, structural information is reserved while data is greatly compressed, and efficient structured query and analysis are achieved. According to the method, the multi-structure sub-modes in the tensor cluster are mined by adopting the Gaussian mixture model, the shared bases are constructed for multi-base projection, and the composite core representation is generated to realize compression and analysis.

The invention relates to the technical field of power system stability control and new energy grid connection, in particular to a method for evaluating the capability of improving the transient stability of a power grid through network construction type energy storage, and the method comprises the steps: collecting data, and obtaining a high-quality data source through time synchronization and quality evaluation; constructing a digital twinborn simulation environment; injecting an optimized excitation signal in a simulation environment, and identifying hidden coupling mode characteristics between virtual inertia control and a power grid oscillation mode by adopting a tensor decomposition method; inputting the features into a deep reinforcement learning agent to explore optimal control parameters, and establishing an agent model by adopting a meta-model optimizer to predict a Pareto frontier; and generating a comprehensive evaluation report containing a transient stability boundary, a risk map, a parameter feasible region and a self-adaptive strategy through high-fidelity verification and calibration of the intelligent agent and the meta-model, and integrating the comprehensive evaluation report to an energy storage control system and a power grid energy management system. According to the method, the problem of stable boundary quantization misalignment caused by implicit coupling is effectively solved.

Techniques and apparatus for improving machine learning model quantification are disclosed. A parameter tensor of a machine learning model is accessed, and a set of rows in the parameter tensor each including one or more outliers is identified. The parametric tensor is decomposed into a first parametric sub-tensor corresponding to the set of rows and a second parametric sub-tensor corresponding to at least one remaining row of the parametric tensor. The first parametric sub-tensor is quantized according to a first quantization scheme, and the second parametric sub-tensor is quantized according to a second quantization scheme. A quantized version of the machine learning model that includes the quantized first parametric sub-tensor and the quantized second parametric sub-tensor is generated.

The application discloses a direct-current system grounding fault safety early warning method based on current characteristics and belongs to the technical field of power system fault diagnosis. The method comprises the following steps: firstly, injecting a safe micro-current disturbance signal into a direct-current system and synchronously collecting bus voltage and branch current response; then, solving a branch coupling admittance matrix at multiple frequency points, performing tensor Tucker decomposition and principal component analysis on the branch coupling admittance matrix, and extracting a first principal component direction vector; next, calculating the variance contribution rate of the vector, combining a dynamic reference and a fault probability accumulator model, and generating a system insulation state index; finally, using a dynamic decision threshold and a confidence positioning mechanism to realize accurate determination and positioning of a grounding fault; the method solves the problems of early warning, anti-interference and complex fault positioning of traditional methods, and improves the safety and reliability of direct-current system operation.

The application relates to the field of data processing, in particular to a tensor database construction method and system of a large model. The method captures tensors and clusters them according to semantics, then uses a Gaussian mixture model (GMM) to mine sub-patterns in the clusters and construct a shared basis. In online processing, a new tensor is decomposed into a composite representation composed of a mode probability and a core tensor through multi-basis projection. The representation retains structural information while greatly compressing data, realizing efficient structured query and analysis. The application mines multiple structural sub-patterns in the tensor clusters through the Gaussian mixture model, constructs a shared basis for multi-basis projection, generates a composite core representation to realize compression and analysis.

The invention provides an influence identification method and system for an enterprise multi-level implicit cooperation relation chain, and relates to the technical field of enterprise management, and the method comprises the steps: obtaining enterprise management data, carrying out the time synchronization modeling between event sequences, and carrying out the association rule mining processing of an abnormal co-occurrence relation, and obtaining an event association initial graph; carrying out processing based on a community discovery algorithm, and carrying out traffic network modeling based on a processing result to obtain an event cooperation association graph; carrying out dynamic graph evolution modeling, and carrying out weak connection sub-graph detection processing on a modeling result to obtain a collaborative chain evolution graph; performing path influence diffusion analysis and state transition probability modeling processing to obtain a cooperative chain influence diffusion network; and carrying out feature analysis and tensor decomposition processing to generate a high-dimensional cooperation risk feature spectrum representing dynamic influence characteristics of the enterprise multi-level recessive cooperation relation chain. According to the method, the non-dominant influence path in the collaboration chain is comprehensively disclosed, and a decision-making auxiliary basis is provided for an enterprise management layer.

The application discloses a method for identifying the scale of a UAV group based on multi-dimensional super-resolution analysis, constructs a distance-azimuth-pitch-Doppler four-order tensor model of a target echo, constructs a four-order tensor decomposition model of the echo signal based on PARAFAC decomposition, performs parameter estimation on the echo signal to obtain parameter information of the azimuth angle, the pitch angle, the speed and the distance of each point target, merges the motion trend of the cluster targets to obtain the aggregation degree of each point target, establishes a cluster target connected graph model to obtain the motion direction corresponding to each group, and geometrically reconstructs the formation shape of the cluster targets to complete the estimation of the number and shape of the UAV cluster formation. The application fully utilizes multi-dimensional feature information of the target to construct a distance-azimuth-pitch-Doppler four-dimensional joint domain, geometrically reconstructs the cluster targets, and realizes the scale estimation and identification of the UAV cluster targets.

The invention relates to the technical field of road construction monitoring, and discloses an asphalt pavement paving quality monitoring system, which comprises a multi-mode sensor array module used for collecting parameter data in a paving process in real time; the data processing module is used for fusing the processed data based on a tensor decomposition algorithm and extracting multi-dimensional features of the paving state; the intelligent analysis module is used for generating a self-adaptive anomaly detection threshold through a time sequence prediction model and identifying an abnormal condition in the paving process; the optimization decision module is used for generating a paving operation parameter adjustment strategy through reinforcement learning based on the abnormal condition; the communication module is used for realizing data transmission and operation instruction issuing between the asphalt paver and the central control system; and the central control system controls the paver to execute operation adjustment operation based on the adjustment strategy. According to the invention, through real-time data acquisition, anomaly detection and adaptive adjustment, the accuracy, intelligence and response speed of paving operation are improved, and the stability of paving quality is ensured.

The invention provides a long-time subway passenger flow prediction method based on deep tensor decomposition and reconstruction, and relates to the technical field of rail transit operation planning. Subway passenger flow data are expressed as three-dimensional tensors of station, time slice and date dimensions, and space and time characteristics of the subway passenger flow data are decoupled and expressed; and initializing corresponding space, time and date potential factors for each dimension, and carrying out nonlinear transformation and feature enhancement on the potential factors through a deep convolutional neural network, so that a complex nonlinear space distribution mode and a time evolution mode of subway passenger flow data can be extracted, and then a reconstruction result is obtained through tensor reconstruction. The constructed tensor decomposition and reconstruction model based on the deep neural network can be used for predicting passenger flow tensor, and finally, a vector autoregression model is introduced to apply time dynamic constraint to date potential factors, so that the model has long-term prediction capability, and accurate and stable prediction of future multi-day (long-term) subway passenger flow is realized.

The invention discloses a traffic data interpolation method based on third-order Tucker decomposition, and the method comprises the steps: firstly constructing a three-dimensional traffic flow tensor according to the features of original traffic spatio-temporal data, so as to comprehensively represent the distribution conditions of traffic flows in space and time dimensions; and then, by constructing a third-order Tucker decomposition model, the third-order kernel tensor is decomposed into three third-order cubic tensors, so that the calculation complexity is effectively reduced, and the spatial-temporal characteristics in the traffic data are accurately extracted. On this basis, an optimization objective function based on third-order Tucker decomposition is established, an efficient solving algorithm is designed based on a least square method, and an approximate tensor is obtained to realize interpolation of a traffic flow missing value. According to the method, the accuracy of traffic flow data interpolation is improved on the whole, the interpolation performance is still excellent under the condition that the data missing rate is high, and the method has good application and popularization values.

The invention discloses a food circulation data analysis method and system based on food big data. The analysis method comprises the following steps of collecting food circulation data, integrating the data through a multi-source heterogeneous data fusion engine, constructing a three-dimensional feature matrix, and eliminating dimensional differences of the data by adopting a tensor decomposition technology. Forming a path topology according to the collected food circulation data, calculating an abnormal confidence coefficient of each logistics node based on a local outlier factor algorithm in combination with a multi-parameter entropy value, and updating the abnormal confidence coefficient in combination with a time sliding window mechanism; establishing a circulation efficiency index, constructing a freshness attenuation model, and performing nonlinear weighting on the circulation efficiency index and a freshness attenuation value based on a fuzzy analytic hierarchy process to generate a circulation quality evaluation value; establishing an elastic supply and demand prediction model, and determining a node optimal inventory distribution strategy; and according to the optimal inventory distribution strategy, adaptive adjustment of the transportation network is realized. According to the invention, the food circulation data is collected and analyzed, so that the food circulation efficiency is improved.

The invention discloses a personalized intelligent recommendation system and method for a financial knowledge platform, and belongs to the technical field of financial systems, and the system comprises a server burying point collection module which collects user key financial behavior event data after business logic processing; the service attribute enhancement module is used for adding a fine service context attribute set to the behavior event; the multi-modal user portrait construction module associates and fuses real-time behaviors and historical business data of the user through a security interface, and generates dynamic user feature vectors from multiple dimensions such as investment specialty and market insight ability; the tensor decomposition recommendation engine is used for modeling the user-article interaction relationship into a four-dimensional tensor containing users, products, specialty and insight, and calculating preference prediction scores through a fusion function to generate a recommendation list; according to the method, on the premise of guaranteeing data security and consistency, personalized recommendation which is highly accurate and conforms to professional ability and preference of financial users is provided for the financial users.

The invention discloses a crowd sensing data recovery method based on three-dimensional convolution and structured information embedding, and aims to solve the problem of sensing data missing caused by node position time-varying characteristics and non-uniform spatial distribution in a crowd sensing network and realize high-quality missing data speculation. The method comprises the following steps: constructing an environment perception tensor; constructing a training set containing a mask tensor; analyzing the training set by adopting a tensor CP decomposition technology to obtain structured feature representation of tensor elements; calculating nonlinear correlation characteristics of tensor elements in the mask tensor through vector outer product operation; inputting the nonlinear correlation features into a preset neural network model to obtain high-order data correlation, accessing a full-connection neural network for mapping, determining predicted values of missing elements in the training set, and obtaining a target training set; and training a preset data speculation model based on the target training set to obtain a target data speculation model.

The invention discloses a tea order digital management method and system based on supply management, and relates to the technical field of order digital management, and the method comprises the steps: carrying out the space-time interpolation processing of pre-obtained picking climate data, generating a climate space-time continuous matrix, carrying out the prediction of a picking period through a phenological prediction model, and obtaining a picking period prediction result; generating picking period data in combination with a pre-acquired picking area; tensor decomposition processing is carried out on order data and picking period data which are obtained in advance to obtain supply and demand collaboration data, and topological data analysis is carried out on the supply and demand collaboration data to obtain a demand topological relation graph; and carrying out demand fluctuation identification on the demand topological relation graph, generating demand fluctuation data, substituting the demand fluctuation data into a sales prediction model to carry out demand trend prediction, and carrying out order dynamic early warning in combination with pre-acquired inventory data. The risk of supply and demand imbalance is effectively reduced, and the limitations of experience driving, passive response and information isolated island in a traditional agricultural supply chain are broken through.

The invention relates to an unlicensed random access unified active device detection method in a high-speed scene, and belongs to the technical field of communication. According to the method, a GF-RA system model is established based on tensor CP decomposition, a Vandermonde structure of a factor matrix is explored and utilized through identical transformation, the ill-conditioned problem caused by large-scale equipment access is solved by adopting a subspace decomposition method, and the accuracy of factor matrix estimation is improved; and the active equipment detection is completed by using the reconstructed factor matrix design by means of an experience threshold and a known training precoder, so that the active equipment detection does not depend on complete channel state information any more and is simultaneously suitable for active and passive application scenarios.

This application proposes a parameter estimation method, apparatus, device, and storage medium, comprising: converting received multi-target echo signals into a multi-target tensor conforming to a unified tensor model of multiple dimensions; denoising the multi-target tensor to obtain a denoised multi-target tensor; inputting the denoised multi-target tensor and VanderMonde constraint information into a trained tensor decomposition model to obtain factor matrices corresponding to each of the multiple dimensions; and estimating the parameters of the multiple factor matrices to obtain parameter sets corresponding to each of the multiple target objects. The embodiments of this application simplify the decomposition method and improve the accuracy of the decomposition by combining VanderMonde constraint information to decompose the tensor.

The application provides a network service quality prediction method based on tensor Tucker decomposition, comprising the following steps: 1, constructing existing network service quality data into a three-dimensional sparse tensor model of user-service-time, used for describing multi-dimensional potential relationship; 2, calculating the similarity between users and the similarity between services respectively by using multiple similarity measurement methods, and constructing a user similarity matrix and a service similarity matrix; 3, clustering and grouping users and services according to the similarity matrix to form a user group set and a service group set, so as to alleviate the data sparsity problem; 4, selecting a target user and a target service, determining the user group and the service group to which the target user and the target service belong, extracting a corresponding sub-data set from the original data and constructing a sub-tensor; 5, performing Tucker decomposition on the sub-tensor, and completing missing data by iterative optimization to obtain a prediction result; and 6, reasonably using an error evaluation index to evaluate the prediction result, so as to verify the model performance.