171 results about "Matrix model" patented technology

The invention discloses an apple anthracnose leaf blight recognition five-dimensional data fusion method based on a patrol robot, relates to the technical field of anthracnose leaf blight recognition, and solves the problems of limited sensing dimension, large illumination interference and inaccurate spatial positioning in the prior art. Comprising the following steps: S1, acquiring surface curvature characteristics of apple leaves, collecting RGB-D images of the leaves, constructing a three-dimensional topological structure of the leaves, and positioning scab space distribution; s2, acquiring spectral reflection information of apple leaves, quantifying biochemical components of chlorophyll a and carotenoid, and constructing a vegetation index feature set sensitive to diseases; s3, apple tree canopy temperature information is collected, a transpiration anomaly detection model is constructed, and multi-modal fusion of thermal infrared and point cloud data is realized; and S4, integrating spatial distribution, spectral reflection and thermal radiation information, constructing a five-dimensional phenotypic characteristic matrix model, and comprehensively analyzing apple phenotypic characteristics. According to the method, the recognition accuracy under the conditions of blade shielding, uneven illumination, blade posture change and environment interference can be remarkably improved.

The invention provides a grid-connected inverter impedance detection method based on an L-shaped filter, and belongs to the technical field of grid-connected inverter impedance detection. A multi-frequency-point disturbance signal is injected into a preset grid-connected inverter working point, time domain data of inverter side current and common coupling point voltage are collected, fast Fourier transform analysis is carried out on the collected data, amplitude and phase information corresponding to each injection frequency point is extracted, and a frequency domain feature vector matrix is established; and establishing a frequency compensation matrix model to compensate and correct high-frequency band measurement data based on a frequency characteristic attenuation vector, calculating to obtain optimal estimated values of power grid impedance and grid-connected inverter output impedance, performing error analysis and precision verification on an impedance detection result, and outputting a verified impedance parameter value. The technical problem of insufficient power grid impedance detection precision caused by high-frequency band attenuation of the L-shaped filter is solved.

The invention discloses a dynamic coupling modeling method and system for an electric power information physical system, and belongs to the technical field of electric power dispatching automation, and the method comprises the steps: analyzing an AGC control process business data flow, and constructing a workflow engine; performing node classification, generating an information node model and an information branch model of the information system, and constructing a rule engine; constructing an information system network matrix model based on the rule engine, and establishing a preset service strategy based on the information system network matrix model; constructing an information physical coupling model of the AGC control process; and according to the AGC actual service, loading a preset service strategy established by the information system network matrix model based on the synergistic effect of the workflow engine and the rule engine, and updating the information physical coupling model. According to the method, a basic model is provided for cross-level closed-loop simulation verification in the AGC process, the problems of unclear information physical dynamic coupling mechanism, insufficient service logic adaptive mapping and model iteration self-updating are solved, and a guarantee is provided for safe and stable operation of a large power grid.

The invention discloses a traffic flow prediction method and device based on a space-time diagram convolutional network, and the method comprises the steps: collecting road traffic flow data, and constructing a traffic congestion index matrix model; calculating a road traffic congestion index based on the collected road traffic flow data, and constructing a traffic flow adjacency matrix by using the road traffic congestion index; performing fast time convolution processing on the traffic flow adjacency matrix to generate a traffic flow data sequence containing spatio-temporal information; inputting the traffic flow data sequence into a space-time synchronization graph convolutional network model STSGCN, further learning a spatial topological structure, and deeply extracting space-time features by stacking multiple layers of space-time synchronization graph convolutional layers; and inputting the spatial-temporal characteristics into a wave loss function-based random vector function link network model Wave-RVFL optimized by an index and trigonometric function optimization algorithm, and predicting the traffic flow. The method can improve the precision and efficiency of traffic flow prediction.

The invention discloses a six-degree-of-freedom mechanical arm disturbance compensation control method based on an iterative learning observer, and the method comprises the steps: taking a visual servo system as a research object, enabling the system to comprise a mechanical arm, a sensing part and a target object, and building a state space model based on a depth-independent Jacobian matrix model with joint speed disturbance; constructing an auxiliary state system and an error transfer equation based on joint sensor information, and designing an iterative learning observer to estimate joint speed disturbance; the disturbance estimation value is embedded into a visual model prediction controller, an optimal control signal is generated, joint speed input is corrected through a feed-forward compensation mechanism, and active disturbance suppression is achieved. According to the method, the tracking response speed of sudden disturbance can be improved, the system stability is maintained under the condition that interference exists in the joint speed, and therefore the visual servo task of tracking the expected trajectory is achieved.

The invention provides a power distribution box insulation monitoring method based on dynamic threshold analysis, and relates to the technical field of power distribution box insulation monitoring, and the method comprises the steps: obtaining multi-class parameter data capable of representing a power distribution box in a healthy insulation state, obtaining a multi-dimensional feature vector set of each time window, and constructing a reference matrix model; inputting a current time window multi-dimensional feature vector set of the distribution box needing to be monitored to obtain an estimated multi-dimensional feature vector set, and calculating a deviation index of the current time window multi-dimensional feature vector set; generating a dynamic early warning threshold value; judging whether a preliminary abnormal signal is triggered or not according to a comparison result of the deviation degree index and a dynamic early warning threshold value; if so, continuously acquiring deviation index slopes of a plurality of later time windows, and performing graded early warning on the insulation state of the distribution box to be monitored according to the deviation index slopes. According to the invention, the insulation state of the distribution box is analyzed based on a working condition self-adaptive dynamic threshold value, and the early and accurate hidden danger identification effect is realized.

The invention discloses a method for determining the position of a box-type transformer of a sand hill photovoltaic power station, and the method comprises the steps: extracting the topographic features of a sand hill through a graph embedding model, converting the topographic features into a high-dimensional vector, and dividing a photovoltaic module layout point matrix in combination with a plane point matrix model layout algorithm. Based on parameters of rated capacity, power supply radius and the like of the box-type transformer, a layout position is determined through the steps of screening effective point locations, dividing power supply partitions, constructing an electrical performance evaluation matrix, optimally screening an optimal candidate point location, performing final verification and the like. According to the method, topographic features and equipment parameters are fused, through multi-step system optimization, the limitation that the prior art depends on experience and single index judgment is overcome, the position determination accuracy and reasonability are improved, the method adapts to the complex sand mountain environment, losses and risks are reduced, and the operation efficiency of a power station is guaranteed.

The invention relates to the technical field of rail transit, and discloses a rail transit tunnel surface environment vibration assessment method, system and device and a medium. The method comprises the following steps: constructing a vehicle-rail-fastener dynamic model; calculating a fastener acting force frequency spectrum according to the vehicle-track-fastener dynamic model; constructing a 2.5 D finite element-basic solution solving model of tunnel-surrounding rock; solving the tunnel-surrounding rock 2.5 D finite element-basic solution solving model to obtain basic solution equivalent vibration source intensity; and calculating earth surface vibration according to the fastener acting force frequency spectrum, the basic solution equivalent vibration source intensity and the field layered soil body power transfer matrix model. According to the method, track vibration, tunnel vibration and site vibration are decoupled, the calculation method with the highest calculation efficiency is adopted in each subsystem, the advantages of each modeling method are fully combined, the calculation efficiency is extremely high, the obtained result is more accurate, the cost is low, and therefore the track traffic tunnel surface environment vibration condition is rapidly and accurately evaluated.

The invention discloses a robot intelligent management and control system for energy conservation and efficiency improvement of a cold end of an indirect cooling auxiliary unit, and belongs to the technical field of intelligent management and control, the robot intelligent management and control system comprises an initial evaluation module, a dynamic planning module and a feedback monitoring module, obtaining a scaling thickness distribution diagram and a scaling coverage rate, and generating a fin flow channel matrix model; the dynamic planning module is used for dividing virtual operation blocks, constructing a scaling planning path in combination with a cleaning index, carrying out rapid rough sweeping, carrying out cavitation early warning, and processing cavitation early warning through a hierarchical control architecture; and the feedback monitoring module is used for constructing a quality deviation set, screening out substandard virtual operation blocks, calculating and correcting priorities, sorting and grading, optimizing and correcting parameters by utilizing a genetic algorithm, and performing feedback regulation and control, so that the cold end heat exchange efficiency is improved to achieve an energy-saving target, a fin structure is protected, and the service life of equipment is prolonged.

The invention discloses a concomitant epilepsy depression onset risk prediction system and device, and the method comprises the steps: constructing an array graph model based on multivariable factors, such as marriage conditions, onset age, epileptic seizure frequency, drug treatment types, SEEG epilepsy origin regions, daily life activity ability assessment scale ADL scores, and the like; through the array graph model, the onset risk of concomitant depression of epilepsy patients can be accurately predicted, so that the diagnosis rate of concomitant depression of epilepsy is improved, and a high-risk patient group of depression can be clinically recognized earlier, so that intervention is performed in time, serious consequences caused by depression onset are avoided in time, and the risk of concomitant depression of epilepsy patients is reduced. And a theoretical basis can be provided for clinically and actively intervening the depressive mood of the epilepsy patient.

The invention discloses a chirp grating design method and device, equipment and a storage medium, and relates to the technical field of optical fibers, and the disclosed chirp grating design method comprises the steps: analyzing a preset target spectrum function to obtain a target grating structure parameter; layering the initial grating into a plurality of sub-grating regions with uniform periods based on the target grating structure parameters, and establishing a layered transmission matrix model; determining a refractive index modulation depth range of each sub-gate region according to the layered transmission matrix model; and performing iterative optimization on grating structure parameters of the initial grating according to the layered transmission matrix model, the target spectral function, the refractive index modulation depth range and a preset error function to obtain a target grating. The performance of the chirp grating can be optimized.

The invention discloses a hospital scientific research consumable purchase and financial system linkage settlement optimization method, and relates to the technical field of purchase management, and the method comprises the steps: collecting scientific research purchase data, building a scientific research purchase data incidence matrix model, and carrying out the automatic verification based on the submission of a consumable application, and calling the scientific research purchase data incidence matrix model, and completing the purchase. After purchasing is completed, a scientific research fund dynamic balance model is constructed to output fund change information; an account structure is monitored and adjusted by outputting fund change information, and scientific research purchase efficiency and risk assessment model output quantitative indexes are constructed. According to the method, the scientific research purchase data incidence matrix model, the scientific research fund dynamic balance model and the purchase efficiency and risk assessment model are constructed, so that real-time linkage of hospital scientific research consumable purchase and a financial system is innovatively realized. The intelligent verification, fund dynamic monitoring and risk quantitative analysis capabilities of the purchasing process are effectively improved, the reasonable fund structure is ensured, and the intelligent level of scientific research management is promoted to be comprehensively improved.

The invention discloses a roadbed compactness rapid detection method and system based on a three-dimensional point cloud technology, and relates to the technical field of three-dimensional scanning. The method comprises the following steps: acquiring normalized point cloud of the inner wall of a roadbed compactness test hole, fitting parameters of a matrix cylinder model by combining macroscopic priori knowledge of a test hole drilling tool and adopting a process fusing two algorithms, and calculating an interior point rate; calculating a normal distance from the point cloud to the surface of the model as a geometric residual error, expanding a cylindrical surface to a two-dimensional plane, mapping the residual error, and performing interpolation to generate an initial residual error grid field; obtaining a filler physical and mechanical parameter derivation constraint condition, and optimizing the initial residual error grid field to obtain a physical compliance residual error grid field; and mapping the data back to three dimensions and superposing the data to a matrix model, generating a triangular mesh model, calculating the volume through numerical integration, outputting a result after interior point rate correction, and calculating the compaction degree in combination with filler performance parameters. According to the invention, the accuracy of roadbed compactness detection is effectively improved.

The invention relates to the field of karst medium modeling and seepage simulation, and provides a three-dimensional reconstruction method and system for a karst coupling medium network. The three-dimensional reconstruction method for the karst coupling medium network comprises the following steps: acquiring corrosion pore characteristics in actual karst geological conditions, and generating an equivalent corrosion pore matrix model; obtaining corrosion hole characteristics in actual karst geological conditions, and generating a three-dimensional corrosion hole equivalent medium model; fracture characteristics in actual karst geological conditions are obtained, an original equivalent three-dimensional fracture network model is generated, fracture parts outside the representation unit body model are cut, and a three-dimensional fracture equivalent medium model is generated; generating a three-dimensional equivalent karst coupling medium network based on the equivalent corrosion pore matrix, the three-dimensional corrosion hole equivalent medium and the three-dimensional fracture equivalent medium model; and performing gridding processing on the three-dimensional equivalent karst coupling medium network to realize fluid mechanics analysis and simulation. The method accurately reflects the equivalent distribution condition and the seepage law of the karst medium.

The invention relates to a digital-intelligent automatic point alignment and full information verification method based on a power distribution terminal tester, which comprises the following steps of: acquiring electrical quantity signals and operating parameters of a terminal through the tester, and establishing communication connection through a digital-intelligent point alignment module; starting a three-remote check matrix model to perform matrix check on telemetering, remote signaling and remote control functions, calling an FFT harmonic synthesis telemetering calibration algorithm to dynamically calibrate a deviation value, introducing an abnormal attribution model based on a Bayesian network to perform probability reasoning on abnormal data, generating a report containing abnormal information and calibration suggestions, and uploading the report. The method improves the verification efficiency and overcomes the problem that the existing mode is tedious; checking accuracy is improved, and the problem that measurement deviation is difficult to correct is solved; abnormality processing capability is enhanced, defects of insufficient attribution capability are made up, stable operation of the power distribution terminal is guaranteed, and support is provided for power grid safety.

The application discloses a two-dimensional woven composite material geometry modeling method, device, equipment and medium, relates to the technical field of composite material structure design and modeling, and is used for solving the technical problems that the simulation model precision and reliability are low in the prior art. The method comprises the following steps: an initial matrix model is created by using a random distribution algorithm and a Boolean subtraction operation; a two-dimensional woven unit cell model is created according to a preset yarn space trajectory formula; a multi-layer two-dimensional woven fiber bundle model is created according to a plurality of two-dimensional woven unit cell models; a reserved fiber bundle matrix model is created by performing a difference set on the initial matrix model and the multi-layer two-dimensional woven fiber bundle model; and a two-dimensional woven three-dimensional model containing random pores is created by in-situ assembly of the reserved multi-layer fiber bundle matrix model and the multi-layer two-dimensional woven fiber bundle model. Therefore, the simulation model precision and reliability can be improved by using the technologies of "random distribution algorithm, Boolean subtraction operation, reserved position and in-situ assembly".

The invention discloses an assembly precision analysis method for a rotary shaft system of a mechanical structure based on a laser tracking measurement system, and in the laser tracking measurement system, the high precision guarantee of an assembled mechanical system is the premise of realizing accurate measurement. In order to obtain the assembly precision of a laser tracking measurement system, the invention provides an assembly precision analysis method based on Jacobi-spinor for a rotary shaft mechanical structure of the laser tracking measurement system. A Jacobian matrix model in the deviation transfer process and a spinor model comprising plane features, cylinder features and plane and cylinder parallel features are combined, a laser tracking measurement system rotating shaft assembly precision analysis model is established, and Monte Carlo simulation is adopted for calculation results to simulate actual assembly errors. A result shows that a difference value between a theoretical assembling condition and an analogue simulation actual assembling condition is 6.79 microns. The result shows that the provided method can qualitatively analyze the assembly precision of the rotating shaft of the laser tracking measurement system.

A weather prediction method includes generating radiance differences as a difference between measured radiances from satellites and forecast satellite radiances generated by a radiative transfer model and forecasted state profiles output by a numerical weather prediction (NWP) model. When the radiance differences exceed a noise threshold, the method includes generating updated state profiles by generating radiance-sensitivities using a Jacobian model and the forecasted state profiles; constructing a Kalman-gain matrix from background error covariance (BEC) matrices and the radiance-sensitivities; generating filtered state-profile changes from the Kalman-gain matrix and the radiance differences; updating the state profiles by adding the filtered state-profile changes to the forecasted state profiles to yield updated state profiles.

The invention discloses a resolution-oriented 4D millimeter wave radar antenna array and physical layout collaborative optimization method, a program, equipment and a storage medium, which are suitable for a radar system with a limited space and a multi-chip structure. The method comprises the steps of firstly collecting system input parameters; then, an antenna design constraint model is constructed, spatial discretization is carried out, and an optimizable matrix model is formed; transmitting and receiving array structures in the azimuth angle direction and the pitch angle direction are jointly designed through a virtual array mapping strategy, so that the requirements of MIMO mapping efficiency maximization and target angle resolution are met. After an antenna structure is determined, an optimization objective function is further constructed based on strong constraint and soft constraint, and an antenna layout scheme which meets the manufacturing requirements and has optimal performance is obtained through a traversal and search method. According to the method, the virtual aperture can be effectively improved, the antenna number redundancy is reduced, the wiring accessibility is optimized, and the resolution and engineering realizability of a 4D radar system are enhanced.

The invention provides a preset bit automatic configuration method and system for a remote intelligent patrol system, and relates to the technical field of intelligent patrol preset bit configuration.The method comprises the steps that a patrol video stream or a video file obtained from the remote intelligent patrol system is decoded, and a target point location detection model and a multi-target tracking algorithm are used for detecting a target point location; identifying each inspection target point location in the video, matching a unique track for the inspection target point location, and generating a point location list containing track information; based on the point location list, constructing a matrix model for describing a coverage relationship between video frames and inspection point locations, solving an optimal frame set covering all inspection point locations through an optimization algorithm, screening out a final optimal preset bit image and corresponding equipment control parameters from the optimal frame set according to a preset rule, and generating a preset bit list. According to the method, automatic identification, trajectory tracking and intelligent optimization of the preset position of the patrol point can be realized, and the configuration efficiency and accuracy of the preset position are remarkably improved.

The application provides a kind of multi-unmanned aerial vehicle target allocation attack-defense strategy game method and system. Multi-unmanned aerial vehicle formation information of first unmanned aerial vehicle formation and second unmanned aerial vehicle formation is acquired to generate strategy pair of multiple unmanned aerial vehicles according to unmanned aerial vehicle information, and strategy group set of multi-unmanned aerial vehicle formation is generated according to strategy pair, strategy pair can include current unmanned aerial vehicle, attack-defense strategy and strategy target;Attack-defense strategy includes attack strategy, avoidance strategy and interference strategy. By calculating the probability distribution of each unmanned aerial vehicle under different attack-defense strategies, a payoff matrix is constructed, and a zero-sum game matrix model is further generated. Based on the zero-sum game matrix model, the attack-defense strategy game solution of multi-unmanned aerial vehicle target allocation is solved. The application fully considers the three attack-defense strategies of unmanned aerial vehicles, and solves the attack-defense strategy game solution, thereby improving the solution quality of unmanned aerial vehicle target allocation problem.

The application discloses a sequence construction method and system. The method constructs a switching matrix based on the number of various types of elements in the sequence, establishes a unified constraint system, and obtains a numerical solution of the switching matrix through optimization. Based on the numerical solution, a graph structure is constructed and connectivity is determined to determine whether a complete sequence can be formed; when the connectivity is satisfied, a target sequence that meets the switching relationship and quantity requirements is generated based on the graph structure. The method supports segment-level structure expression, modeling of the head-tail relationship of the augmented matrix, and sub-sequence combination expression, and is suitable for multiple sequence construction tasks. Accordingly, the application also provides a sequence construction system, which includes a matrix modeling module, a constraint construction module, a solving module, a connectivity determination module, and a sequence generation module.

The application relates to a seepage simulation method in a multi-scale fractured medium based on coupling finite elements, which comprises the following steps: 1. acquiring statistical distribution characteristics of related parameters of multi-scale fractures in a simulation domain, and constructing a multi-scale fracture geometric model; 2. respectively performing intersection line detection on each scale fracture in the multi-scale fracture geometric model; 3. simplifying the multi-scale fracture geometric model; 4. adopting coupling units to establish coupling relationships among the each scale fracture models; and 5. simulating a multi-scale discrete fracture-matrix model. The application is based on a coupling finite element method, and couples three scale fracture conceptual models, so that the difficulties in calculation precision, calculation cost and mesh partitioning of a single traditional model in solving multi-scale problems are avoided, and a feasible new way is provided for large-scale complex rock mass underground water flow analysis.

The application provides a scientific research business management and control system based on large model technology, comprising a data acquisition module for acquiring multi-source heterogeneous data in the whole process of power scientific research business; a knowledge extraction module for extracting knowledge from the multi-source heterogeneous data according to a power knowledge graph and a data extraction model to obtain a semantic matrix; a model reasoning module for reasoning knowledge from the semantic matrix according to an intelligent reasoning large model to obtain decision information; a bidirectional association interface for determining a reasoning direction according to a structured field constraint of the data extraction model and determining prior knowledge of the intelligent reasoning large model according to the power knowledge graph; and a business management and control module for managing and controlling the scientific research business according to the decision information. Through multi-module cooperation and deep integration of large models and domain knowledge, the application realizes full mining and accurate analysis of multi-source heterogeneous data in the whole process of power scientific research, makes the management and control basis more comprehensive, and guarantees the accuracy and scene adaptability of the reasoning direction.

The application discloses an online network measurement flow recovery method, a terminal device and a storage medium. Network flow data at each moment is modeled according to a two-dimensional matrix model; a basic subspace column is trained according to flow data at historical moments, and indexes of the columns are marked, and a leverage score is obtained by calculation; according to the indexes of the subspace columns, a subspace column that needs full sampling at the current moment is selected, and only part of elements in the remaining columns are sampled; whether sampling needs to be continued is judged according to a sampling matrix at the current moment; when a maximum residual value is less than a certain threshold, unsampled data is recovered by using subspace matrix filling; when the maximum residual value is greater than a certain threshold, columns that need to be continuously measured are selected according to residual sizes, a column space is updated, and then the unsampled data is recovered. The application solves the problems of low recovery precision and inability to recover online in the prior art, and realizes online network measurement with low overhead.

The present application relates to a kind of multi-energy flow carbon footprint tracking method based on carbon memory matrix and conservation constraint, belong to the technical field of energy internet, industrial carbon management and intelligent manufacturing.The multi-source energy-carbon coupling data pedestal is constructed, the unified data of multiple time scales is realized;The energy storage carbon memory matrix model containing time dimension and temperature stratification is established, the dynamic evolution process of carbon intensity in energy storage system is described;On this basis, the multi-energy flow directed carbon flow chart is constructed, the unified carbon tracking and apportioning of electric, heat, cold multi-energy flow are realized;Carbon flow conservation constraint and weighted least squares state estimation method are introduced, the physical consistency verification and self-repair of carbon data are realized;Finally, the mixed integer linear programming model of explicit embedded carbon memory constraint is constructed, the optimal multi-energy collaborative scheduling of whole life cycle carbon emission is realized.The present application realizes the change from static accounting to dynamic tracing of carbon footprint, from post-event statistics to active optimization, significantly improves the reliability of carbon data and low-carbon scheduling effect.

The invention discloses a network communication anomaly detection method and device, equipment and a storage medium, and the method comprises the steps: constructing a real-time communication matrix and a historical communication matrix, carrying out the initialization setting of the real-time communication matrix and the historical communication matrix, continuously capturing a network message at a protocol stack bottom layer of the equipment, and carrying out the analysis of the network message, thereby constructing a quintuple, and carrying out the detection of the network communication anomaly. Performing Hash operation on the quintuple by using a preset Hash function, performing matrix updating on the real-time communication matrix according to a Hash operation result, performing matrix comparison on the real-time communication matrix and the historical communication matrix according to a preset period to obtain a difference communication matrix, performing matrix evaluation on the difference communication matrix, and performing matrix calculation on the difference communication matrix according to a matrix evaluation result. And judging whether the current network communication is abnormal or not. According to the method, all communication links of the equipment can be monitored in real time by constructing the network communication matrix model of the global port, and the defect of one-sidedness of a traditional detection method is overcome.

The application discloses a kind of direct current motor iterative learning control optimization methods for executing change task, it is related to direct current motor control field.The method is based on the closed loop feedback control system of direct current motor and parallelly joins iterative learning controller, based on promotion technique will direct current motor control system be converted into time series input-output matrix model.Under the norm optimization framework, optimal iterative learning control algorithm is designed, through the combination of batch-to-batch repeated learning and batch real-time feedback, so that the system basically realizes zero error tracking to expected output.Based on the optimal input sequence and error sequence obtained by repeatedly executing a task, the feedback and feedforward controller are integrated into a new learning-based feedback controller using least squares fitting method, and finally it is applied to the system executing the change trajectory task.This method transfers the historical learning experience to a new task without limiting its time length, and realizes the trajectory tracking of direct current motor change task without relearning.

The application provides a manufacturing cloud service recommendation generation method, comprising: obtaining manufacturing service descriptive information and label information of a cloud platform; using an LDA topic model to vectorize and model the manufacturing service descriptive information, obtaining p topic information of the manufacturing service, and obtaining p*q-dimensional service topic word vectors through Word2vec mapping expansion; inputting all the word vectors into CNN to learn the deep features of the manufacturing service; further calculating the manufacturing service similarity matrix through the deep feature similarity; using an improved K-Means clustering to cluster the manufacturing service label information to obtain the manufacturing service category, further obtaining a user category matrix model and a service category matrix model, and obtaining a coefficient to construct a user preference matrix; and combining the manufacturing service similarity matrix and the user preference matrix to obtain a manufacturing service recommendation list. The application solves the problems of incomplete manufacturing service information feature extraction and sparse user service matrix in the prior art.

The application discloses a dynamic time-sensitive application network routing method and device, a storage medium and an electronic device, and the method comprises the following steps: in the case that a node application initiates a data request to a center node, the center node determines a candidate path based on a network state matrix model, wherein the node application is a user node mapped on a data plane of a logical network topology of a dynamic time-sensitive application network, and the center node is located on a control plane of the logical network topology; a communication time delay of each candidate path calculated based on a routing metric is used to determine a target path from the candidate path, and a path planning parameter is formed; on the basis of the same time reference, the center node performs quadratic programming on the target path, and forms a shaping configuration parameter; and the center node issues the path planning parameter and the shaping configuration parameter to the node application, so that the node application performs data forwarding based on the target path. The application solves the technical problem of deterministic time delay transmission of time-sensitive applications in a dynamic complex network.