183 results about "Real-time web" patented technology

The invention discloses a network security defense method and system based on incremental network attack analysis learning. The method comprises the following steps: collecting initial network flow data, and extracting a feature vector; and collecting real-time network flow data, performing segmentation processing based on a sliding time window, extracting time sequence association features from the segmented data, and matching the time sequence association features with the feature library to identify potential attacks or abnormal behaviors. And when the time sequence correlation feature is not matched with the feature library, marking the time sequence correlation feature as a candidate novel attack feature, calculating a mahalanobis distance between the time sequence correlation feature and a known attack feature to determine the attack variability, and dynamically adjusting the weight of the time sequence correlation feature. And inputting the adjusted feature weight and the real-time flow feature into a deep reinforcement learning detection model, and generating and updating a network security defense strategy. The scheme of the invention can effectively identify novel attack behaviors, dynamically respond and optimize defense strategies, and improve network security.

The invention relates to an information transmission method and device and a storage medium, and the method comprises the steps: S1, obtaining a predicted value of information transmission efficiency based on real-time network bandwidth data; s2, based on the predicted value of the information transmission efficiency, in combination with a real-time network load, carrying out capacity segmentation adjustment and sending priority configuration of the data packet, and generating data packet transmission configuration; s3, calculating data packet transmission time based on data packet transmission configuration in combination with a real-time network bandwidth, and determining a compression level of the data packet; s4, generating a hierarchical compression data packet based on a compression hierarchical calculation result; and S5, based on the hierarchical compressed data packets, performing error detection and correction in the transmission process, and performing data integrity verification on the transmitted data packets to generate an information transmission verification result. Compared with the prior art, the method has the advantages of reducing data transmission time, improving network utilization rate and the like.

The invention relates to the field of network security, and discloses a problem line determination method based on IP address classification and identification, which comprises the following steps of: collecting public network IP behavior data in a network flow log, extracting multi-dimensional characteristics such as a time regularity entropy value, a protocol diversity proportion, target port dispersity and a request rate change rate, and determining a problem line according to the extracted multi-dimensional characteristics; constructing an IP association graph fusing feature similarity, time synchronism and physical position constraints; an abnormal IP group is identified by adopting a community division algorithm, and the detection sensitivity is adaptively optimized according to a real-time network load and a historical false alarm rate in combination with a dynamic threshold adjustment mechanism based on a PID feedback control model; the attack type is judged through protocol-port mapping and Mahalanobis distance statistics, alarm information containing an abnormal IP list, a time window and a classification result is generated, and the alarm information is linked with safety equipment to execute a defense strategy. According to the method, the defects of a traditional method in the aspects of multi-dimensional attack recognition, dynamic environment adaptation and cooperative attack detection are overcome.

The invention relates to a network security vulnerability intelligent detection method based on artificial intelligence and big data, and the method comprises the steps: carrying out the dynamic time synchronization processing of multi-modal network security data, and generating a multi-source data flow with aligned time sequences through cross-modal correlation analysis; extracting cross-modal features from the data stream, and performing semantic fusion on the cross-modal features in combination with a vulnerability knowledge graph to generate a multi-dimensional feature vector; training the multi-dimensional feature vector through a hybrid model to obtain a vulnerability detection model, detecting real-time network behavior data by using the model, and outputting a vulnerability probability and an abnormal risk score; and performing automatic vulnerability verification according to the vulnerability probability and the abnormal risk score to obtain a verification result, and updating the vulnerability detection model according to the result. The system can effectively improve the accuracy, timeliness and stability of network security vulnerability detection and reduce the false report and missing report rate through multi-modal data collaboration, hybrid model dual detection and closed-loop optimization mechanisms.

The invention discloses a software query information management method and system, and belongs to the technical field of software information security, and the method specifically comprises the following steps: after a query request of a user is received, creating a dynamic topology network architecture according to semantic features of query contents; creating a virtual query container corresponding to the dynamic topological network in the isolated memory space, wherein the virtual query container generates a dynamic encryption channel bound with the current topological structure; dynamically embedding user permission labels into topological network nodes through an attribute-based encryption algorithm, and establishing a mapping relation between permission characteristics and network paths; when distributed query is executed in the virtual query container, the topological node connection weight is dynamically adjusted according to the real-time network load; before a query result is fed back to a user, data elements are recombined based on a reverse entropy value of an original topology network architecture; and when an abnormal access feature is detected, triggering a topological network structure dynamic failure mechanism.

The invention relates to the technical field of audio playing optimization, and discloses a low-delay multi-room audio playing optimization method and system, a storage medium and equipment, and the method comprises the steps: building a global time synchronization system based on a precision clock protocol in a playing equipment cluster, and enabling all playing equipment to operate based on a unified time reference; monitoring network state parameters in real time, dynamically calculating and adjusting the audio buffer depth of each playing device based on the monitored network quality, and obtaining audio data in advance according to the playing progress and the network state; generating a global playing timestamp according to the global time synchronization system, and controlling all playing devices to perform synchronous starting and progress control of audio playing based on the global playing timestamp; a transmission strategy of audio data is dynamically optimized based on real-time network state parameters, packet loss recovery and bandwidth adaptation are achieved, sub-millisecond multi-room audio synchronous playing is achieved, and meanwhile performance parameters can be automatically optimized according to different network environments and hardware configurations.

The invention discloses an eSIM service cost intelligent optimization method based on different operator information sources, and particularly relates to the technical field of operator network management. The method comprises the following steps: collecting eSIM tariff rules of a plurality of operators and real-time charging data of eSIM terminals, extracting tariff terms and conditional expense trigger mechanisms hidden in charging behaviors of different operators by using an association rule mining algorithm, and constructing a multi-level cost structure model; on the basis of a network registration behavior log of the eSIM terminal, predicting the probability that each operator signal source has an abnormal registration event, and evaluating the risk that each operator signal source triggers a hidden cost trap in combination with a multi-level cost structure model; and according to the real-time network signal parameters of the eSIM terminal, a signal source recommendation strategy is generated in combination with the risk assessment result, and the eSIM terminal is driven to execute a signal source switching operation, so that real-time adjustment and cost optimization control of communication network selection are realized, and the service stability of the eSIM in a cross-network use scene is improved.

The invention discloses a zero-trust data access control method and system and a storage medium, and the method comprises the steps: carrying out the identity authentication of a user logging in a client, and obtaining the current terminal equipment information of the client, the current network information and the historical access information of the user after the authentication succeeds, the trust level of the authorization request and the current access permission matched with the trust level are obtained; acquiring a real-time access action of a user and real-time terminal equipment information and real-time network information of a client to perform threat detection on the authorization request, and adjusting the current access permission so as to generate a real-time access strategy corresponding to the client according to the adjusted real-time access permission; and acquiring a data access request uploaded by the client according to the real-time access strategy, acquiring access data required by the client according to the access request, and sending the encrypted access data to the client, thereby improving the security of data access.

The invention discloses a deep learning-driven multi-base-station radio map rapid construction method and system. The method comprises the steps of generating a radio map original data set of a current environment; processing the radio map original data set to obtain a radio map sample data set; and constructing a neural network model, and training the neural network model according to the radio map sample data set to obtain a radio map construction model. According to the method, the learning ability and generalization performance of the model are enhanced through residual connection and batch normalization by adopting a double-path neural network architecture, so that comprehensive processing of signals of multiple base stations and rapid and high-precision radio map prediction are realized, and the requirements of real-time network planning and optimization are met. Furthermore, by processing the radio map original data set, each pixel data in the image includes path gain information, thereby providing an accurate signal propagation path and improving identification accuracy.

The invention discloses an intelligent heterogeneous network fusion optimization method and device based on a large model, and the method comprises the steps: collecting real-time network state data, service demand data and user equipment data of various heterogeneous networks, and carrying out the preprocessing; key features capable of reflecting real-time changes of network states and service requirements are extracted from the preprocessed data; a self-adaptive multi-modal fusion algorithm is adopted, modal characteristics of heterogeneous network data are automatically identified, and weights and parameters in the fusion process are dynamically adjusted, so that the optimal fusion effect is achieved; building a large model driven by deep learning in combination with a cross-network context awareness optimization strategy; and the intelligent fusion module dynamically selects the optimal network path and transmission strategy according to the output result of the large model. According to the method and the device, intelligent fusion of the network is realized by adopting a self-adaptive multi-modal fusion algorithm and a cross-network context sensing optimization strategy based on a large model of deep learning.

The invention relates to a large language model reasoning enhancement method and device based on a multi-source retrieval and summary module, and the method comprises the steps: taking Qwen2.5-32B as a large language model, and carrying out the fine adjustment of the large language model through a preset retrieval data set; multiple rewards are introduced through reinforcement learning to improve the self-cognition ability of the large language model; constructing a vector database and a real-time network retrieval system based on the FAISS, and performing multi-source retrieval on the text data through the vector database and the real-time network retrieval system; text information of the text data is summarized through a preset small language model, and the text information is output to the large language model. According to the method, the self-cognition ability of the big language model is improved by adopting reinforcement learning (GPRO), and the multi-source retrieval module is constructed, so that the data with similar semantics can be quickly queried to assist the big language model in deeply understanding the semantic connotation of the data, and the accuracy and rationality of retrieval are improved.

The invention discloses a hydrological data synchronization method and device based on distributed storage and a medium, and the method comprises the steps: dividing hydrological data collected by different hydrological monitoring point sensor devices in real time into a plurality of different grades according to the data emergency degree, and adding a space-time label to the hydrological data; fragmenting the hydrological data based on the space-time dimension to obtain a standard data block, and determining a regional storage node corresponding to the standard data block; according to the real-time network state, determining an optimal transmission path corresponding to different levels of hydrological data; and based on the consistency transmission rules corresponding to the multiple different levels, synchronizing the standard data blocks to the corresponding regional storage nodes through the optimal transmission path. Through level-to-level management and intelligent routing decision making of hydrological data, an optimal transmission path is selected according to characteristics of different levels of data and network states, resource waste and transmission delay caused by a traditional unified synchronization strategy are avoided, and the data synchronization efficiency is remarkably improved.

The invention relates to the technical field of data processing, in particular to a vCPE resource allocation method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining system operation data of each monitoring point, inputting the system operation data into a preset anomaly detection model, outputting an abnormal data point, determining a vCPE resource allocation strategy according to the abnormal data point, inputting the system operation data into a preset target prediction model, and obtaining a vCPE resource allocation result. And outputting predicted operation data in a future time period, adjusting a resource allocation strategy according to the data, and adjusting vCPE resource configuration according to the adjusted strategy. According to the method and the device, the limitation that traditional vCPE resource allocation is based on static configuration or simple rule scheduling is broken through, the resource allocation strategy can be dynamically adjusted according to the real-time network condition and the user service requirement, the vCPE resource utilization rate is effectively improved, the network performance is optimized, and the problems of low resource utilization rate and limited network performance caused by an existing resource allocation mode are solved.

The invention discloses a power Internet of Things data transmission method and device and a medium, and belongs to the field of power network networking, and the transmission method comprises the steps: collecting power data such as current, voltage and power in real time based on a sensor, and carrying out the de-noising of the data through Kalman filtering; layering packaging is carried out according to the data type, a data compression algorithm is applied to compress the data, the transmission bandwidth is reduced, the compressed data is encrypted, and integrity verification information is added; dynamically selecting a transmission channel by using a reinforcement learning algorithm according to a real-time network state; the data are packaged and sent to the cloud through the selected transmission channel, the data are received at the cloud, and the power data are stored and managed by using a time sequence database; and analyzing the received data, generating a real-time report, carrying out anomaly detection and load prediction through a random forest algorithm, and carrying out adaptive adjustment according to the real-time monitoring data. By dynamically selecting a transmission channel and a self-adaptive transmission strategy, network instability is dealt with, and timely transmission of data is ensured.

The invention provides a dynamic honey point deployment method and system based on an attack graph and a Shapley value. The method comprises the following steps: scanning and exploring a protected network every first preset time; an attack graph corresponding to the current network environment of the protected network is constructed according to the scanning and probing result, the probability that each vulnerability in each edge in the attack graph is utilized is calculated based on CSVV evaluation, and the attack difficulty of each edge is defined; according to the attack difficulty of each edge, the implementation probability of an attack path is calculated and obtained, and the expected loss of each attack path in the attack graph is obtained; and constructing a game model, solving a Shapley value of each edge based on a Monte Carlo method, and carrying out edge sorting on all the edges in the attack graph according to the Shapley value of each edge so as to generate an optimal honey point deployment strategy according to an edge sorting result. According to the invention, the honey point deployment strategy can be dynamically adjusted according to the real-time network security situation in the attack and defense process.

The invention discloses a storage resource allocation method and device, electronic equipment and a storage medium, and relates to the technical field of resource scheduling, and the method comprises the steps: determining the priority weight of each storage volume based on network state information and performance index information at intervals of a preset time period or if the network state information represents that network congestion occurs at present, allocating a corresponding bandwidth to each storage volume based on the priority weight; if the current to-be-processed task does not need to be operated across the heterogeneous storage device, determining a target transmission path and a target storage volume of the current to-be-processed task according to the priority of the current to-be-processed task and the bandwidth corresponding to each storage volume; and processing the current to-be-processed task based on the target transmission path and the target storage volume. The technical problem of storage performance jittering caused by the fact that the storage resource distribution cannot be dynamically adjusted according to the real-time network state is solved, and the technical effects of reducing the sudden delay or lagging phenomenon and improving the system stability are achieved.

The embodiment of the invention discloses a video safety playing method and device based on multi-dimensional protection and a medium, and relates to the technical field of video transmission, and the method comprises the steps: obtaining terminal device information under the triggering of a video playing request of a target playing terminal, performing two-factor authentication on the target broadcast terminal based on the terminal equipment information and terminal registration information pre-stored in a block chain node, and determining an equipment authentication result; collecting a real-time network index of each transmission path through a pre-deployed network probe, and screening a target transmission path from the plurality of transmission paths based on the real-time network indexes; and when the device authentication result is passed, acquiring watermark fragment data corresponding to a target video stream in a pre-constructed block chain node, so as to transmit the watermark fragment data to the target broadcast terminal through the target transmission path, and monitoring the transmission process in real time to realize safe broadcast of the target video stream.

The invention discloses a power production OTN resource scheduling method based on a particle swarm optimization algorithm, and the method comprises the following steps: 1) constructing a digital twinborn model of a power production OTN network through a digital twinborn modeling method of multi-class model fusion; and 2) modeling is carried out on a power production OTN resource scheduling problem, a swarm intelligence optimization method based on a particle swarm optimization algorithm is adopted, a neighborhood search strategy and a global search strategy suitable for power production OTN resource scheduling are designed, and the power production OTN resource scheduling method based on the particle swarm optimization algorithm is realized. The improved particle swarm optimization algorithm is adopted to dynamically adjust the position and speed of the particles according to the real-time network state, self-adaptive optimization is achieved, the method adapts to the changing network environment, multiple optimization targets can be considered at the same time, the scheduling scheme is comprehensively optimized, and the method has higher global search ability, high convergence speed and high robustness.

The invention relates to the technical field of industrial networks, in particular to a time-sensitive service flow resource adaptation method for large-bandwidth multicast flow awareness, which comprises the following steps: S1, a controller collects registered service flow information and real-time network state information; s2, after service flow priorities are divided according to the information, a mapping relation between service flows and multiple queues is established, and a forwarding rule is established according to the mapping relation; s3, establishing an optimization target and a resource constraint according to the mapping relation and the forwarding rule; and S4, designing a Markov decision according to the optimization target and the resource constraint, training the neural network model according to the Markov decision and the double-delay depth deterministic strategy gradient so as to output a scheduling decision, and obtaining a multi-dimensional resource adaptation result by a solver according to the scheduling decision. According to the invention, a basic queue model support is provided for end-to-end deterministic transmission of the unicast stream and the multicast stream through the mapping relation and the forwarding rule, the utilization rate of network resources and the number of service streams capable of being subjected to incremental scheduling are improved, and the worst end-to-end delay of the large-bandwidth multicast stream is reduced.

The present application relates to the technical field of network anomaly monitoring, and discloses a network anomaly monitoring method and system for a switch, the method comprising: collecting historical network flow data of a network switch and preprocessing it to obtain preprocessed network flow data; performing multi-scale feature extraction and feature dimension reduction on the preprocessed network flow data to obtain a fused feature vector; training a switch-level model and a central model based on the fused feature vector; performing anomaly detection on real-time network flow data through the switch-level model to obtain preliminary anomaly detection results, and performing in-depth analysis in combination with the central model to obtain in-depth anomaly detection results; performing spatiotemporal correlation analysis and multi-source data fusion on the preliminary anomaly detection results and the in-depth anomaly detection results to obtain a comprehensive anomaly analysis report; generating an adaptive defense strategy based on the comprehensive anomaly analysis report, and executing the adaptive defense strategy in the network switch, thereby improving the accuracy of network anomaly monitoring of the switch.

The invention provides a network equipment user identification method and system based on network protocol analysis, and relates to the technical field of network equipment user identification, and the method comprises the steps: collecting network protocol data generated when a user interacts with network equipment; performing protocol analysis on the network protocol data to obtain protocol features; according to the protocol characteristics, user category labeling is carried out on the network protocol data; performing differential coding and splicing on the protocol features to form a network protocol feature vector; constructing a network equipment user identification model based on a deep neural network, and predicting a user category according to the network protocol feature vector; comparing the user category prediction result with the user category labeling result, and performing iterative training on the network equipment user identification model; collecting real-time network protocol data generated when a user interacts with network equipment, and converting the real-time network protocol data into a real-time network protocol feature vector; and identifying unknown user categories through the trained network equipment user identification model.

The invention provides a lightweight satellite network security management system and method. According to the system, unified management of a satellite network security control plane is realized through the centralized network controller. The system has the functions of satellite node security authentication, dynamic security policy generation and issuing, real-time network monitoring, rapid security event response and the like. The network controller comprises an authentication module, a strategy generation module, a monitoring module, an event response module and the like, a security strategy can be customized according to satellite characteristics, and security threats can be dealt with through real-time monitoring and timely adjustment. The invention also comprises a lightweight design which is particularly suitable for a satellite environment with limited resources, and a method for continuously optimizing a security policy based on monitoring data and event records. In addition, the provided user interface allows a satellite network administrator to conveniently perform policy configuration and event processing. According to the system, through centralized management and dynamic adjustment, the safety and the operation efficiency of the satellite network are remarkably improved, the system has the advantages of light weight, flexibility, real-time performance and the like, and the safety management requirement of the modern satellite network is met.

The invention discloses a flow filtering method and device for extracting quintuple flow characteristics, and the method comprises the steps: obtaining a quintuple characteristic sequence of a historical network communication data sample; in a plurality of time windows, extracting joint probability distribution among different quintuple features in the quintuple feature sequence to obtain time sequence correlation features; constructing a communication prompt word according to the time sequence association feature; inputting the communication prompt word into a preset large language model for training to obtain a target large language model; performing anomaly analysis on the real-time network communication data through the target large language model to obtain an analysis result; determining an exception type based on an analysis result; and determining a traffic filtering strategy according to the exception type and a preset strategy library, and performing traffic filtering through the traffic filtering strategy. According to the invention, the accuracy of abnormal traffic filtering can be improved.

The invention provides a hybrid congestion control method and system based on dynamic adjustment of a monitoring time period, and relates to the technical field of communication, and the method comprises the steps: obtaining a real-time network state, and calculating a round-trip delay change according to the network state to obtain network state data; performing normalization processing on the network state data to standardize the network state data, comparing the standardized network state data with a preset threshold to classify the network state data after normalization processing, and calculating a monitoring time period value according to a classification result to dynamically adjust the monitoring time period value; judging whether the current time reaches a monitoring time period value or not; and if yes, switching the current congestion window to a congestion window which is generated based on deep reinforcement learning and is adjusted by an optimal action according to the monitoring time period value. Compared with a traditional fixed monitoring time period scheme, the method not only has higher environment generalization ability, but also can quickly adapt to network condition changes through real-time state classification.

The invention belongs to the technical field of internet online conference scheduling, and particularly relates to an internet-based online conference scheduling system and method. According to the method, a space-time weight mechanism ensures that a high-priority scheme is generated in one step, multiple rounds of back-and-forth communication are reduced, overlapping and absent risks in a selected time period are effectively avoided through historical behavior prediction and conflict verification, an interactive feedback mechanism quickly responds to newly-added conflicts, the stability of the scheme is improved, an audio and video transmission strategy is dynamically reserved, and the user experience is improved. Resource allocation and historical QoS data driving are intelligently adjusted according to real-time network conditions, delay, lagging and interruption caused by network problems in a conference can be remarkably reduced, in combination with personalized prediction of a topological structure, an internal cooperation habit, a visual time selection topological graph and an online negotiation tool are better matched and organized, transparency and sense of participation are enhanced, and the network quality is improved. And the multi-dimensional conflict and the space-time weight are fused, so that the balance optimization of the comprehensive cost and the efficiency is realized.

The invention provides a power distribution network evaluation method and system, and relates to the technical field of power system operation control, and the method comprises the steps: obtaining an extreme event index of a power distribution network in a current time period according to a measurement data flow, and judging whether an extreme event exists in the power distribution network or not according to the extreme event index; the measurement data streams are classified through a data quality grading mechanism, and then an observation weight matrix of the extreme events is constructed; supplementing missing data for the missing data segment according to the observation weight matrix, dynamically updating topological information of the power distribution network in combination with the supplemented missing data, and obtaining a real-time network model corresponding to the extreme event according to the topological information; and according to the real-time network model and the extreme event index, obtaining a state evaluation result of the power distribution network in the current time period. According to the invention, the evaluation precision of the power distribution network under extreme conditions is improved.

The invention discloses a multifunctional AR glasses and mobile phone interaction method and device, and relates to the field of equipment interaction. In the method, a plurality of communication channels are established between an AR glasses end and a mobile phone end through a bidirectional protocol, a protocol encapsulation layer is overlaid on the plurality of communication channels, and a mapping relation between a data type and the plurality of communication channels is established; mapping the user operation into a cross-device instruction stream, and selecting an optimal channel according to the data type and the mapping relationship of the cross-device instruction stream; determining a compression and encryption strategy according to the data type and the real-time network condition, processing the cross-device instruction stream by using the compression and encryption strategy, and transmitting the processed cross-device instruction stream through the optimal channel; and obtaining interaction data and an equipment state between the AR glasses end and the mobile phone end, and when the interaction data or the equipment state is abnormal, switching the optimal channel to the target channel. By implementing the technical scheme provided by the invention, the interaction efficiency of the AR glasses and the mobile phone is improved, and the data transmission reliability is ensured.

The invention discloses a cross-platform advertisement putting strategy collaborative management method and system, and relates to the technical field of advertisement putting, and the method comprises the steps: uploading a plurality of advertisement test samples to a putting platform test library for testing, obtaining real-time network information, and carrying out the evaluation of a test result, selecting the advertisement test sample with the highest advertisement putting return rate data as a main push advertisement, setting an advertisement putting strategy according to the hot time period of the platform, then pre-putting the main push advertisement on each platform according to the advertisement putting strategy to evaluate the pre-putting effect, establishing a revenue prediction model according to a deep reinforcement learning algorithm, and predicting the revenue of the main push advertisement. The method comprises the steps of predicting the income of advertising on each platform in real time, dynamically distributing a budget weight according to the predicted income data of the advertising on each platform to obtain an advertising strategy distribution table, and formulating an advertising income distribution contract according to a preset advertisement click rate and a conversion value. And according to the advertisement revenue distribution contract, calculating and distributing revenue to an account corresponding to the platform.

The invention relates to the field of network public opinion data monitoring and analysis, in particular to a clustering analysis-based network public opinion data monitoring and processing method, which comprises the following steps of: establishing multi-stage data characteristics of network public opinion data by utilizing real-time network public opinion data; acquiring a basic classification result of the real-time network public opinion data based on clustering analysis by utilizing the multi-stage data features of the network public opinion data; according to the method, a basic classification result of real-time network public opinion data is obtained, a network public opinion data monitoring processing result is obtained by utilizing a basic classification result of the real-time network public opinion data, historical semantic features and network communication features of real-time data are respectively output into independent comparison data on the basis of clustering analysis, result misjudgment of a single semantic or single source IP is avoided first choice, and then through an iterative optimization process, the real-time network public opinion data monitoring processing result is obtained. According to the scheme, the complete process of implementing the scheme is improved, the monitoring efficiency is improved, the clustering analysis database is gradually expanded according to the scheme implementation time, and the generalization ability of scheme implementation and the processing accuracy of large-batch data are improved.

The invention discloses a network flow prediction method based on a decoupling dynamic space-time diagram neural network D2STGNN. The method comprises the following steps: constructing a network flow data processing and representing module, and collecting and preprocessing flow data; then, a prediction model is constructed based on D2STGNN, a space-time framework is decoupled to decompose a network traffic space-time signal into a node-traffic association signal and a network-traffic time trend signal, and by means of a space-time local convolution layer, a gating circulation unit and a multi-head self-attention layer, an inter-node association process and a capture time sequence dependency relationship are respectively modeled; and meanwhile, a time attenuation mechanism is introduced, and the real-time dynamic capturing capability of the model on the network flow is enhanced in combination with adaptive attenuation parameter adjustment and a layered attenuation mechanism. And finally, predicting real-time network flow data by using the trained model, and evaluating a prediction effect through related evaluation indexes. According to the method, the change trend of the network flow can be efficiently and accurately predicted, and powerful support is provided for reasonable distribution of network resources, network performance optimization and the like.