3532 results about "Streaming data" patented technology

The present invention relates to the technical field of big data processing, and discloses a multi-source heterogeneous data fusion and processing method based on big data, including the following steps: data preprocessing, cleaning, complementing and feature extraction of multi-source heterogeneous data, providing standardized and high-quality input for subsequent processing; dynamic knowledge graph modeling, constructing a dynamic knowledge graph based on the semantic relationship of multi-source data, and generating a unified semantic representation; cross-modal alignment and deep fusion, semantically aligning multi-modal data, and generating a unified comprehensive representation through deep fusion of features; real-time stream processing, dynamically updating and processing real-time updated stream data to achieve real-time fusion of data; adaptive optimization, dynamically adjusting fusion strategy parameters, and improving the adaptability and robustness of the fusion strategy. The present invention realizes efficient fusion, real-time processing and semantic consistency expression of multi-source heterogeneous data through dynamic knowledge graph modeling, cross-modal alignment, real-time stream processing and adaptive optimization.

The invention provides an AI-based factory abnormal behavior identification monitoring method and system. The method comprises the steps of obtaining continuous video stream data collected by a plurality of monitoring devices in a target factory in real time; performing time dimension segmentation and space dimension preprocessing on the continuous video stream data to obtain a multi-frame behavior feature set corresponding to the dynamic behavior sequence, and inputting the multi-frame behavior feature set into a pre-trained abnormal behavior recognition model, performing cross-modal association fusion on the skeleton key point features, the environment associated object contour features and the action trajectory features to generate a target behavior feature vector, performing similarity matching with a preset abnormal behavior feature library, determining an abnormal behavior category, and triggering a corresponding early warning signal according to the behavior type when it is determined that the target behavior feature vector belongs to the abnormal behavior category. And real-time alarm display is carried out. According to the invention, the problem that the existing monitoring system is insufficient in accuracy and real-time performance can be solved.

Systems and methods are described for a implementing a streaming data processing system that includes a pool of pre-configured resources and a pool of dedicated resources. The streaming data processing system can implement a processing pipeline using compute resources. The pool of pre-configured resources can support previews of processing pipelines for a plurality of users and the pool of dedicated resources can support full deployments of processing pipelines for a particular user. The streaming data processing system can implement a preview of a processing pipeline using a pre-configured resource of the pool of pre-configured resources. Further, the streaming data processing system can implement the processing pipeline using a dedicated resource of the pool of dedicated resources. The streaming data processing system can provision the dedicated resource and deploy the processing pipeline using the dedicated resource.

The invention relates to the technical field of intelligent control, provides a flow regulating valve servo force control method and system based on a force sensor, and aims to solve the technical problems of response delay, weak overshoot suppression capability and poor long-term operation stability. The method comprises the following steps: acquiring a servo driving current data set and a valve displacement track data set of a target regulating valve; performing pressure feature mapping processing on the servo driving current data set to generate a pressure fluctuation feature set corresponding to the driving current waveform data; the pressure fluctuation characteristic set and the valve displacement track data set are input into a preset force control decision model for dynamic matching processing, and a servo control instruction set is generated; executing multi-stage dynamic adjustment operation on a servo driving unit of the target adjusting valve according to the servo control instruction set, and generating real-time pressure balance state data; and iteratively updating the dynamic matching processing parameters of the force control decision model based on the deviation value of the real-time pressure balance state data and the preset pressure reference value.

This disclosure relates to methods, systems, and devices for AI / ML assisted management of a radio access network (RAN). In particular, an LLM is employed to generate answer to user queries, for example in the form of code that can be executed upon streaming data from the RAN, including performance management and control management databases. To provide prompts the LLM can understand, an LLM agent employs a knowledge base and experience base to transform the language of the queries from the highly technical telecommunications domain to a general domain. Frequently asked questions and their answers may be stored in the experience base and used instead of using the LLM.

The invention provides a real-time map updating method and system based on multi-source geographic information data fusion. Wherein a three-dimensional model is constructed by integrating traffic flow data, remote sensing images and road network topology, the road traffic pressure is quantified, and a feature correlation index is established. And obtaining a vehicle displacement vector and a speed gradient in combination with laser point cloud and video monitoring data, generating a dynamic road feature data set, and establishing a mapping relationship with the index table. On the basis, a road state prediction model is constructed, and through correlation analysis of traffic pressure data and a real-time feature data set, deep correlation between vehicle motion features and road states is established. And dynamically adjusting a road network connection structure, synchronously updating mapping parameters and predicting model precision, and realizing real-time adaptive adjustment of the road network topology. According to the technical scheme provided by the invention, the real-time performance of map updating and the road condition prediction accuracy are remarkably improved, and the urban traffic congestion index can be reduced.

The invention relates to the field of personnel action monitoring and recognition, and discloses an intelligent human shape trajectory prediction and alarm system and method based on multi-modal video analysis, and the system comprises a video collection module which is used for collecting video stream data of a monitoring scene; the data processing module is used for extracting motion state information of the target human shape; the trajectory modeling module is used for constructing a trajectory model of the target person; the risk assessment module is used for calculating a behavior risk value; the alarm module is used for triggering alarm; and the optimization module is used for optimizing the trajectory model. Target motion state information is extracted through the data processing module, high-precision input is provided for the trajectory modeling module, and the trajectory modeling module generates a prediction trajectory in combination with a dynamic optimization algorithm. Compared with a scheme depending on coarse-grained detection or simple position tracking in the prior art, the method has the advantages that fine modeling and dynamic prediction of the motion state are realized, the problem that complex behaviors cannot be accurately reflected is solved, and meanwhile, the track prediction precision and the system response capability are improved.

The invention discloses an intelligent financial risk early warning method and system based on a management decision, and relates to the technical field of data intelligence, and the method comprises a multi-source heterogeneous data collection module which obtains enterprise financial data, supply chain data, market public opinion data and industry reference data in real time through an API interface, risk keywords are extracted from news, social media and policy documents through a natural language processing technology according to the market public opinion data; the streaming data processing engine is constructed based on an Apache Flink framework, performs windowing processing on the real-time data stream, calculates the dynamic fluctuation ratio of financial indexes by sliding a time window, and compares the dynamic fluctuation ratio with a preset industry risk threshold value; and a risk decision fusion model, a dynamic threshold adaptive module and a man-machine collaborative early warning terminal. According to the method, millisecond-level financial index fluctuation monitoring is realized through a streaming computing framework, a knowledge graph and a natural language processing technology are fused, a risk entity and a causal chain are extracted from unstructured data, and a multi-dimensional risk portrait is constructed.

The invention relates to the technical field of data processing, and discloses a PTP network performance monitoring method and system based on big data analysis, and a medium. The method comprises the following steps: acquiring PTP message data, equipment state data and RTP stream data; preprocessing the data and calculating a clock error; constructing a time sequence database and an analysis model; real-time monitoring and anomaly detection are executed; performing performance scoring and fault positioning by applying machine learning; and carrying out network optimization and predictive maintenance. According to the PTP network performance monitoring method based on big data analysis, global state monitoring, intelligent anomaly detection, accurate fault positioning and predictive maintenance of the PTP network are achieved, and therefore the high-precision time synchronization requirement of a broadcast and television IP making and broadcasting system is guaranteed.

The invention discloses a highway situation awareness and accident risk early warning method fusing multi-source data. The method comprises the following steps: S1, constructing multi-source data fusion and multi-scale traffic flow data; s2, studying and judging an accident influence factor mechanism, and excavating factors influencing the occurrence of a traffic accident; s3, early warning of accident risks is carried out, and the safety risk level of the road section is evaluated; and S4, researching the beforehand management and control system, and constructing a beforehand management and control strategy system based on deep mining analysis of traffic accidents and actual operation experience. According to the invention, on the basis of the high-speed portal data and the charging flow data, the detection data of the thunder-vision integrated roadside equipment are fused for fusion processing, so that fine traffic operation parameters such as traffic flow and vehicle speed before and after a historical accident are obtained; traffic flow characteristics causing accidents are mined, and a traffic accident risk dynamic early warning management and control model and a post-accident evaluation mechanism are established based on the traffic flow characteristics, so that a management department is helped to recognize potential risks in advance.

The invention provides a large-scale scene multi-level-of-detail cloud rendering processing method and device based on 3DGS, and relates to the technical field of three-dimensional modeling, and the method comprises the steps: dividing a target modeling scene into a plurality of sub-blocks; performing particle redundancy reconstruction and overlapping region marking on boundary regions between adjacent sub-blocks of each sub-block to obtain processed sub-blocks; performing multi-detail level division on each processing sub-block to generate a particle level set; determining a current visual area according to the user motion data, and scheduling a target hierarchy of a particle hierarchy set in the current visual area; the rendering tasks of all the processing sub-blocks of the target hierarchy are distributed to a plurality of rendering nodes to execute real-time rendering operation; and performing video stream coding on pictures rendered by each rendering node, decoding and displaying received video stream data, and performing particle level updating and re-rendering operation according to an interaction instruction. According to the invention, high-quality detail rendering can be realized for a model of a large scene.

The invention relates to a nonlinear system reduced-order model generation method fusing physical constraint and digital twinning, and belongs to the technical field of digital twinning. The method comprises the steps of performing sparse sampling on high-dimensional dynamic response data, extracting a dominant mode of high-fidelity model data, and retaining a mode of a physical abrupt change region through physical field gradient constraint in the extraction process; the high-fidelity finite element model and the low-fidelity simplified model are combined, and a mixed order reduction framework based on a multi-fidelity agent model is constructed; fusing a data driving error and a physical mechanism residual error in data-physical joint driving model training and optimization based on a mixed order reduction framework, and updating a primary function of a dominant mode based on online stream data; and performing cross validation on the output of the high-fidelity finite element model and the reduced-order model, and improving the precision of the reduced-order model through an iterative optimization strategy. The intelligent order reduction problem of the model is realized, and the accuracy and rapidity of the order reduction model are ensured.

The invention relates to the technical field of data processing, and discloses a comprehensive management method and system for a smart park, and a storage medium. The method comprises the steps that environment, energy consumption and people flow data are collected through the Internet of Things, and a basic data set is obtained through edge node processing and semantic alignment; constructing a digital twin system by using a spatio-temporal data cube and a graphical mode; constructing a land digital identity system by adopting space-time coding and block chain technologies; performing anomaly identification and risk analysis on the activity data to obtain supervision early warning data; performing multi-dimensional analysis on the enterprise operation data to form an enterprise portrait; and presenting resource distribution through mixed reality, and analyzing cross-department data to obtain operation efficiency data. According to the invention, holographic perception, intelligent supervision and cooperative service of park elements are realized.

The invention discloses a computer big data information processing system, which comprises a data acquisition layer, a data processing layer and a data processing layer, wherein the data acquisition layer is used for accessing structured, unstructured and streaming data by using a multi-source adapter and Apache NiFi, executing format standardization, and extracting basic metadata and semantic tags through a rule engine and an NLP model; the metadata intelligent management layer integrates four modules, namely a federal learning framework for realizing cross-domain dynamic classification labels, an intelligent contract for real-time uplink storage evidence blood relationship change, a Neo4j combined graph neural network for constructing a knowledge graph for mining implicit association, and a reinforcement learning engine for optimizing a storage strategy based on frequency and risk indexes; the distributed storage calculation layer is used for processing batch and real-time metadata by adopting a Cassander + MinIO mixed framework and Spark / Flink, and dynamic partition balance performance is realized; and the application service layer is used for outputting functions of blood relationship query, classified browsing, compliance report and the like through a Vue.js portal and a Spring Cloud micro-service API (Application Program Interface) to form a full-link closed loop.

The invention provides a traffic participant behavior analysis-based edge calculation signal lamp control method and system, and the method comprises the steps: obtaining video stream data of a target intersection, carrying out the frame-by-frame behavior state analysis of the video stream data through a spatial-temporal feature coding network, extracting the behavior state features of each traffic participant, and carrying out the analysis of the behavior state features of each traffic participant; and inputting the behavior state characteristics into a preset behavior matching model, generating a behavior triggering identifier of the traffic participant in the current time window, generating a signal lamp control instruction of an edge computing node according to the time sequence relevance between the behavior triggering identifier and the phase state of the current signal lamp, and sending the signal lamp control instruction to the edge computing node. And finally, a signal lamp phase switching time sequence of the target intersection is adjusted based on the signal lamp control instruction, so that the traffic participants meeting the traffic rule conflict condition obtain the traffic priority under the signal lamp phase switching time sequence. According to the method, the traffic accident risk is reduced, and meanwhile, the overall traffic efficiency of the intersection is optimized by dynamically adjusting the minimum response period and the phase duration parameters.

The invention provides an asset library model security risk situation assessment system and method, and the system comprises a data collection and processing subsystem, a multilayer asset association graph construction module which constructs the function dependence, data flow dependence and physical connection relation of assets in a resource library based on a graph database; the risk propagation probability calculation module is used for simulating a conduction path of an attack from an edge node to a core resource by adopting a Monte Carlo method; and the time sensitivity analysis module is used for dynamically adjusting threat weights in different time periods through a sliding window mechanism. According to the method, real-time comprehensive data is provided for evaluation through a multi-source streaming data fusion engine and distributed edge computing nodes; accurately analyzing asset association and risk propagation by applying a multi-layer asset association map and a Monte Carlo method; by means of a reinforcement learning decision framework and incremental model updating, self-adaptive decision optimization is achieved, and novel attacks are coped with in time; based on a visual interface and a closed-loop feedback control loop, the visual effect and the evaluation accuracy are improved.

The invention relates to the technical field of page rendering, and discloses a progressive page rendering method and system based on streaming data, and the method comprises the steps: carrying out the data segmentation through obtaining a data stream, user interaction data and equipment performance parameters, and obtaining data blocks; then, performing cache management in combination with the data, and constructing a multi-level cache pool; thirdly, performing priority grading and sorting on the data blocks to form a rendering task queue; and according to the equipment performance parameters, optimizing a task sequence and obtaining an optimized task sequence. Next, combining the optimized task sequence and user interaction data, predicting data about to enter a viewport, and generating a viewport pre-rendering task; and finally, according to the viewport pre-rendering task, the multi-level cache pool and the equipment performance parameters, performing rendering strategy optimization to obtain a dynamically adjusted rendering task flow. The method can realize dynamic resource scheduling.

The invention relates to the technical field of electrical engineering, and provides an electric energy storage system optimization scheduling method and system based on reinforcement learning, so as to solve the problems that it is difficult to effectively evaluate and control the battery aging risk in the frequent charging and discharging process and it is difficult to meet the real-time scheduling requirement in the edge side low computing power environment. The method comprises the steps of obtaining battery charging state and health state data in an electric energy storage system and power grid load and power grid electricity price data of an industrial and commercial park, and generating a state input data set; using the reinforcement learning agent to obtain a charging and discharging scheduling instruction; performing power distribution on the electric energy storage system, the multi-port converter power interface and the park load to generate multi-source cooperative power flow data; in combination with historical operation data, a reinforcement learning agent is utilized to optimize the charging and discharging scheduling instruction; and performing optimal scheduling of the electric power energy storage system according to the optimized instruction. According to the invention, efficient, intelligent and adaptive optimization scheduling of the power energy storage system in the industrial and commercial park is realized.

The invention relates to the technical field of e-commerce, in particular to an e-commerce promotion method and system based on cloud computing, and the method comprises the following steps: building a user behavior sequence through log collection, extracting a sudden change point through dynamic time warping, generating a dynamic parameter through segmented integral, calculating a difference degree through gray correlation analysis, and reconstructing a weight set. And flexibly scheduling advertisement resource distribution, and generating a delivery instruction. According to the method, user click stream data is collected, a behavior sequence model is constructed, continuous dynamic characteristics of user browsing tracks are captured, time sequence relevance is analyzed, fluctuation peak points of user behavior paths are aligned, interest mutation nodes are captured, abnormal behavior modes are recognized, and implicit preferences and explicit feedback parameters in the interaction process are quantified. Establishing an associated feature analysis mechanism, resetting advertisement weight parameters, matching real-time requirements of users with putting strategies, mapping weights to distributed advertisement nodes, dynamically scheduling computing resources, adapting display requirements and resource allocation, and forming a closed-loop optimization link.

The invention relates to an optimization processing method for a streaming data dynamic window, and the method comprises the steps: taking streaming data as a dynamic system, and calculating the data disturbance energy at each moment on a time axis; constructing a window trigger point according to the potential energy curve; when the trigger potential locally reaches a peak value or the trend changes, adaptively dividing the window; the starting boundary and the ending boundary of each window are defined as elastic intervals; before the window slides or is closed, window boundary self-adjustment and complementation are carried out according to compensatory analysis of subsequent data streams; taking a large window as a container, and dividing sub-segments according to data density change; partial states are shared among the sub-windows, and fine-grained analysis including anomaly detection and instantaneous peak value recognition is executed; introducing a life cycle termination relationship driven by an entropy threshold: when the window activeness is lower than the entropy threshold, triggering state compression or early release; the accuracy and dynamics of window division are realized, the response capability of a system to data fluctuation and burst is enhanced, and the elasticity and fault-tolerant capability of window boundaries are improved.

The invention discloses a financial risk assessment method based on big data, and relates to the technical field of finance, and the method comprises the following steps: S1, obtaining structured data, unstructured data and real-time streaming data of a target entity through a multi-source heterogeneous data collection module; s2, constructing an association relationship graph, and modeling risk propagation paths of a target entity and associated nodes thereof based on a graph neural network; s3, performing feature alignment and joint representation learning on the structured data, the unstructured text data and the time series data through a multi-modal data fusion module; and S4, based on the causal inference model, separating causal features and hybrid variables of the target entity risk event, generating causal risk factors, quantifying risk infection paths between nodes by setting an enterprise guarantee network and a supply chain relation graph dynamically constructed in a graph neural network, effectively identifying hidden risk nodes, and improving the risk assessment efficiency. And the chain reaction risk caused by the default of the associated enterprise is reduced.

The invention relates to the technical field of industrial monitoring, in particular to a fan running state comprehensive monitoring system and a noise and current detection method thereof, and the system comprises a noise data acquisition module, a vibration noise correlation analysis module, a running current acquisition module, an abnormal current detection module, an equipment fault diagnosis module and a performance degradation prediction module. According to the method, the sound data are collected, the frequency and amplitude changes are analyzed, the types of various sound sources are recognized, abnormity caused by mechanical parts is found in the early stage, association analysis of vibration data and sound data is combined, understanding of the physical state of the fan is enhanced, and by monitoring current data and the fan blade rotating speed in real time, the fan blade rotating speed can be accurately detected. Abnormal detection of the current data and evaluation of the current stability are realized, a data basis is provided for fault diagnosis, the operation efficiency and the performance degradation trend of the fan are evaluated in combination with the current data, the service life and the maintenance period of equipment are effectively predicted, and the operation strategy and the maintenance plan of the fan are optimized.

The invention discloses a storage security protection and fire-fighting early warning method based on an edge AI mode, and belongs to the technical field of safety monitoring, and the early warning method employs an edge AI artificial intelligence video data processing analysis method to carry out intelligent security protection behavior identification early warning and intelligent fire-fighting behavior identification early warning; the edge AI artificial intelligence video data processing analysis method comprises the steps that a data center pushes a trained model to an edge AI server, and the edge AI server analyzes streaming data collected by edge Internet of Things equipment in real time and carries out AI reasoning; the edge AI server uploads the received new condition related data to the cloud to train the model, and the trained model is pushed to the edge AI server again; according to the invention, security and protection and fire-fighting applications are deeply fused, an information island is thoroughly broken, data sharing and interconnection are realized, unified output of early warning signals is ensured, and the emergency response speed and accuracy are improved.

The invention discloses a multi-source modal data fusion-oriented charging pile fault diagnosis method and system. The method comprises the following steps of: acquiring message stream data, audio data and infrared sensing data associated with a charging equipment fault through a sensor system; and based on the characteristics of each modal data, fault characterization and identification are carried out by adopting a random forest, a convolutional neural network and a YOLO-v8 target detection model. According to the marked fault data, a mapping relation among multi-source data is established, and a Dempster-Shafer multi-source fusion theory is utilized to aggregate a multi-modal fault judgment result, so that the equipment fault is accurately diagnosed. According to the method, the accuracy of fault diagnosis and diversified presentation of fault types are effectively improved, the repair rate is reduced, and the method has a good engineering application prospect.

The invention belongs to the technical field of intelligent traffic, and particularly discloses a traffic jam analysis method based on an intelligent traffic platform, which comprises the following steps: collecting multi-source traffic flow data of a target road section in real time, judging whether the target road section is jammed in combination with historical data, verifying the traffic flow in a continuous monitoring period to improve the accuracy, and if the target road section is jammed, judging whether the target road section is jammed or not. If yes, the congestion influence range is dynamically determined by identifying a core area and analyzing the upstream and downstream speed propagation trend, then frequent or accidental congestion types are accurately distinguished and the congestion level is evaluated by calculating the deviation degree of current data and a historical traffic mode in multiple dimensions, and finally the congestion influence range is determined according to the congestion types, the congestion level and the increase trend. A differentiated traffic dispersion scheme is generated and executed; according to the invention, through deep fusion of type identification, degree evaluation, range determination and scheme generation links, an integrated decision link is formed, and the intelligent level and response efficiency of the system for coping with a complex congestion scene are significantly improved.

The invention relates to the technical field of traffic prediction, in particular to a multi-layer space-time structure correlation perception traffic flow prediction model, which comprises a space-time structure decomposition layer for decomposing original traffic flow data into a road hierarchical structure feature matrix, a dynamic time period feature matrix and a hidden space topology feature matrix; the dynamic adjacency matrix generation layer is used for constructing a self-adaptive adjacency weight matrix based on the hidden space topological feature matrix; the cross-level correlation perception layer performs bidirectional feature modulation on the road hierarchical structure feature matrix and the dynamic time period feature matrix to generate a space-time coupling feature tensor; and the prediction layer inputs the space-time coupling feature tensor into a space-time diagram convolution prediction network and outputs a traffic flow prediction value in a future time period. According to the method, the expression ability of the model on potential heterogeneous association between the nodes is improved, and the generalization ability across regions and time periods is effectively improved.

The invention discloses a campus face recognition abnormal behavior monitoring method and system based on artificial intelligence, and the method comprises the steps: collecting the video stream data of a person in a campus in real time, and recognizing the behavior characteristics of the person in a video frame in real time through employing a lightweight target detection algorithm; extracting facial features of the personnel by using a pre-trained face recognition model, and recognizing identity information of the personnel; generating a personnel behavior semantic description vector with a spatio-temporal context based on the behavior characteristics and the identity information; inputting the personnel behavior semantic description vector into a dynamic early warning threshold engine, and outputting an abnormal behavior probability and risk level evaluation result; and triggering an early warning information pushing mechanism according to the abnormal behavior probability and the risk level evaluation result, and sending early warning information to a corresponding responsible person terminal. By using the embodiment of the invention, the accurate and automatic association of the abnormal behavior and the personnel identity can be realized, and the early warning accuracy, the handling response speed and the system adaptive ability of campus safety monitoring are improved.

Systems and methods are described for detecting anomalies within time series data using one or more online machine learning models. In one example, time series data may be obtained that spans a moving time period of a set length, to be used as training data. A machine learning model may be trained using the training data to identify anomalies in the time series data by upon detecting an estimation of an anomalous value in the time series data, pausing input of additional values of the training data into the machine learning model, and resuming input of the additional values of the training data into the machine learning model based on detecting at least one non-anomalous value. The trained machine learning model may then be used to detect a first anomaly in additional time series data by comparing the at least one additional value to the trained machine learning model.

The invention discloses a method and a system for monitoring energy conversion carbon measurement based on dynamic carbon monitoring. The method comprises the following steps: acquiring energy flow data, carbon emission parameters, equipment working conditions and environment data in real time; constructing a carbon emission correction model containing a dynamic energy conversion coefficient and real-time fuel components; adopting an efficiency compensation function and a working condition self-adaptive correction function related to an equipment load rate; and visually displaying the real-time numerical value of the carbon emission and triggering the early warning of the carbon emission based on the model output. The system comprises a multi-source data real-time acquisition module, a dynamic carbon measurement model and a real-time monitoring and early warning platform, the method has the advantages of accurate decision support, environmental supervision enhancement and adaptation to energy transformation.

The invention relates to the technical field of security and protection management, in particular to a campus security and protection management system based on deep learning, which improves the accuracy and robustness of identity recognition by acquiring access control card numbers, face images or fingerprint features and generating standardized identity authentication data. And on the basis of a comparison result of the identity authentication data and the campus database, a behavior chain initialization identifier is generated, and accurate identity binding of the school entering personnel is realized. Furthermore, by collecting multi-camera image stream data, pedestrian re-identification and similarity calculation are executed by using a deep feature matching network, and a cross-camera continuous trajectory data set is generated. And matching the behavior track data set with the conventional path template to generate a behavior offset feature vector. And carrying out joint modeling on the behavior offset characteristics and the identity information through a graph neural network model containing an attention mechanism, and outputting a behavior purpose label and a risk grade score. And a graded security response instruction is generated based on the risk score, so that the missing report rate and the false report rate are effectively reduced.