56 results about "Causal graph" patented technology

The invention discloses a causal evaluation method for advertisement putting. The method comprises the following steps: acquiring a plurality of data sets of advertisements put on a take-out platform by merchants operating on the take-out platform; inputting the plurality of data sets into a causal graph mining model set to obtain a first candidate causal graph set; filtering the first candidate causal graph set based on an expert knowledge base to obtain a second candidate causal graph set; obtaining the weight of the path according to the relationship between the node and the path of each second candidate causal graph in the second candidate causal graph set, and aggregating the plurality of second candidate causal graphs based on the weight of the path to obtain a causal assessment large graph; and based on the causal evaluation large graph, evaluating the causal relationship between the advertising parameters and the effect indexes of the merchant advertising. According to the method, the accuracy of advertisement effect evaluation can be improved through accurate causal relationship analysis, the advertisement putting efficiency can be improved, and the putting strategy can be optimized.

An analytics system identifies interventions for individual samples from a set of samples with a mixture of interventions. Given a causal graph, a set of baseline samples, and a set of samples with interventions, a set of intervention tuples is determined that represents the mixture of interventions for the set of samples with interventions. Each intervention tuple in the set of intervention tuples identifies an intervention and a mixing coefficient representing a percentage of samples with the intervention. An iterative process is used in which a set of intervention tuples is determined for N variables and then lifted to a set of intervention tuples for N+1 variables until all variables from the causal graph have been considered, providing a final set of intervention tuples. The final set of intervention tuples is used to match individual samples from the set of samples with interventions to interventions.

This invention provides a time-causal graph-based fault diagnosis method for carrier-based aircraft in landing scenarios, addressing the technical problem that data-driven carrier-based aircraft fault diagnosis methods cannot meet the high-precision requirements of practical fault diagnosis. First, the basic structure and variable parameter table of the triplet time-causal graph model are defined, transforming the nonlinear mathematical model into a small-deviation form. The time-causal graph model of the carrier-based aircraft is determined using an equation method. Typical failure modes are classified according to the residual characteristics of the failure injection parameters, ultimately obtaining a failure feature vector table for subsequent fault diagnosis. Then, the residuals of the observable parameters of the carrier-based aircraft under fault conditions are calculated. Finally, the original failure feature vectors inferred from the time-causal graph model are corrected using the failure feature vectors trained from simulation data, resulting in a feature vector table of failure modes trained from data, thus completing the quantitative description of fault features. This method achieves carrier-based aircraft fault diagnosis based on the time-causal graph model in landing scenarios that meets the high-precision requirements of practical fault diagnosis.

The present application belongs to the field of intelligent medical treatment, and particularly relates to a sepsis early warning method based on a causal chain alignment lightweight large language model, comprising obtaining clinical time series data of a sepsis patient in a period before the onset of the disease, and constructing a causal graph according to the characteristics of the data of each patient; based on medical knowledge, weighting two nodes with an edge relationship in the causal graph; converting the clinical data of a patient to be diagnosed into natural language text, and screening abnormal data therefrom, taking the abnormal data as a root node to screen a causal chain from the causal graph; taking the natural language text and the causal chain of the patient as inputs of a lightweight large language model, judging whether the patient will have a sepsis attack within N hours in the future, and giving a reasoning chain. The present application effectively improves the stability and accuracy of sepsis early warning on the basis of maintaining the advantages of low computational overhead and easy deployment in a real ICU environment of the lightweight model.

This invention provides a computing resource scheduling method and system, belonging to the field of computing resource scheduling technology. The scheduling method includes: collecting multi-source heterogeneous data to generate a full-dimensional fused dataset; using a pre-defined unsupervised learning technique to transform the full-dimensional fused dataset into a standardized unsupervised feature vector set; unifying the dispersed multi-dimensional potential energy components and using a pre-defined unsupervised clustering algorithm to perform comprehensive calculations of potential energy values; selecting unsupervised auxiliary causal variables and causal relationships to generate a preliminary system causal graph, and using a pre-defined quantization method to calculate the causal effect strength of the preliminary system causal graph; transforming the target task of the target computing facility into quantifiable sub-objectives, and using the do-calculus operation of the system causal graph model to generate multiple sets of candidate parameter tuning schemes. This invention achieves a balance between global optimization and local efficiency, proactively avoids potential risks, dynamically balances multiple objectives and continuously evolves, and maximizes overall benefits.

This application provides a fault propagation prediction method, apparatus, and related equipment, belonging to the field of information technology. An embodiment of this application provides a fault propagation prediction method, comprising: acquiring real-time operation and maintenance data of multiple nodes and historical operation and maintenance data within a preset first time range; constructing a service causal graph based on the historical operation and maintenance data and a preset topology constraint matrix, wherein the topology constraint matrix is ​​used to determine potential causal relationships between the multiple nodes; adjusting the edge weights of the causal edges of the faulty nodes according to the real-time operation and maintenance data to obtain an updated service causal graph, wherein the faulty nodes are one or more nodes among the multiple nodes that have experienced a fault; and predicting the fault propagation probability of the fault among the multiple nodes based on the updated service causal graph. This fault propagation prediction method can improve the accuracy of fault propagation prediction.

This application discloses a reliable AI scheduling method, system, and computer equipment for microgrids. It includes: conditional independence testing of multi-source data; constructing a candidate causal graph among variables, constrained by prior knowledge in the power sector; calculating the differences in causal effects caused by fluctuations in the current scheduling strategy; performing counterfactual reasoning verification based on different differentiated scheduling strategies; and scoring the current scheduling strategy on causal interpretability based on causal effect differences, counterfactual verification results, and link integrity, followed by AI scheduling. This application reveals the causal interaction mechanism between variables, and the counterfactual reasoning that changes in variables lead to changes in scheduling results, improving the interpretability of the scheduling strategy. It can trace the causal driving factors of each scheduling instruction, significantly reducing the rate of human intervention. Prior knowledge in the power sector ensures safe and compliant scheduling, making the results more reliable. It solves the problem of dependence on human intervention in the application of AI scheduling in microgrids.

The application relates to the technical field of industry chain scheduling, and discloses an industry chain adaptive scheduling method and system fusing deep reinforcement learning, which comprises the following steps: obtaining multiple variables in a target industry chain in real time, and constructing a dynamic causal graph; predicting a potential risk area based on causal entropy, and judging whether a strategy evolution is triggered; when the strategy evolution is triggered, generating a new strategy paradigm according to instability information and activating the new strategy paradigm; interacting the new strategy paradigm with a simulation environment of the target industry chain, determining an effective scheduling instruction, and deploying the effective scheduling instruction to a physical execution system of the target industry chain. The application breaks through the limitation of traditional lag performance indicators by constructing a dynamic causal graph containing causal strength in real time, monitoring stability by means of causal entropy, predicting potential risks, generating a new strategy paradigm according to instability information after triggering the strategy evolution, determining a scheduling instruction through simulation interaction and deployment, realizing early identification and forward prediction of causal instability, and dynamically adapting to the logic change of the industry chain.

This invention relates to the field of fault diagnosis technology, specifically disclosing a cross-operating-condition fault diagnosis method based on a stable causal graph convolutional domain generalization network. The method acquires multi-channel observation data of equipment under different operating conditions and explicitly introduces operating condition variables to label the corresponding operating conditions. Based on the multi-channel observation data and operating condition variables, a cross-operating-condition stable causal graph is constructed. The causal strength between adjacent observation variables in the cross-operating-condition stable causal graph is quantified to generate a causal strength matrix. Using the causal strength matrix as the graph topology weight, the multi-channel observation data is input as node features into a graph convolutional network for feature extraction to obtain graph-level embedding vectors. A diagnostic model including a classification branch and a domain generalization branch is constructed and trained. The trained diagnostic model is used to diagnose faults in equipment data under the target operating condition. This invention improves the generalization performance and fault diagnosis accuracy of the diagnostic model when facing unknown operating conditions.

The application relates to a causal graph neural network prediction method and medium for eliminating AD false correlation. The method first constructs a graph structure of SNPs and brain images, extracts features using a time series graph network and maps the features to a disease quasi-time axis, and then divides stages and generates a graph sequence. The causal contribution degree between nodes is calculated through a time series prediction network to obtain a causal prior graph; direction correction is performed by using counterfactual flipping and graph convolution to obtain a refined causal graph. The features are subjected to causal enhancement, and double-mode information is fused through cross-attention. Finally, a classifier is input to realize AD staging prediction. The method overcomes the limitation that existing methods are difficult to distinguish between causality and false correlation in cross-sectional data, and through explicit modeling of disease time evolution and causal structure, the ability to capture early Alzheimer's disease and subtle pathological patterns is enhanced, thereby improving the accuracy, interpretability and cross-domain generalization ability of classification prediction.

This invention relates to the field of distributed data link monitoring technology, and discloses a method and system for monitoring and handling anomalies across the entire data link based on intelligent operation and maintenance. The method includes: acquiring multi-source operation and maintenance data and performing cross-link time normalization; identifying abnormal events based on a unified operation and maintenance feature sequence; constructing a delay causal graph carrying propagation delay based on the abnormal event sequence; performing counterfactual dynamic Bayesian inference based on the delay causal graph; and executing processing actions. Compared to existing technologies that rely on static thresholds for anomaly detection, especially under conditions of slow SQL, connection pool queuing, and mutual propagation of interface latency, this invention addresses the technical problem of failing to identify cross-component anomaly propagation relationships and achieve secure closed-loop processing. By identifying counterfactual root causes through anomaly event recognition and selecting security constraint processing actions, this application achieves accurate localization of the root causes of anomalies across the entire data link, improving the accuracy of anomaly response and operational stability of the distributed data link.

The application relates to an abnormality diagnosis method and storage medium for a double-membrane process water plant, and relates to the field of intelligent operation and maintenance of water plants. The method comprises the following steps: collecting multi-source sensor operation data of the double-membrane process water plant; performing effective data screening on each monitoring variable constituting a multivariate time series within a preset time window to obtain a time window level abnormality score of each monitoring variable; under the prior constraint of a process flow, an initial directed graph is constructed according to the process flow, and the monitoring variables are classified to obtain a directed causal graph for representing the double-membrane process structure; a strategy for searching upstream from an abnormality node of a water production state to an upstream root cause node on the directed causal graph is obtained through reinforcement learning training; when it is detected that the water production state variable is abnormal, the strategy is called to search for an abnormal link from a downstream abnormality node to an upstream root cause node on the directed causal graph, and a diagnosis result is output. The application solves the technical problems of rapid positioning of a double-membrane water plant full-process abnormality and traceability of an interpretable link level root cause.

This invention relates to the field of fault location technology, and discloses a method and system for rapid fault location in operation and maintenance. The method includes: acquiring operational data and denoising it to obtain a denoised data sequence; segmenting the sequence and identifying anomalies, determining the anomaly range through trend offset, and obtaining anomaly markers; if the waveform distortion metric of the anomaly marker exceeds a threshold, obtaining the anomaly location through time series analysis; extracting periodic sequences, grouping similar behaviors among devices to obtain associated device clusters; constructing a causal graph among devices, obtaining a list of potential anomalous devices through propagation path backtracking and node dependency sorting; assessing the impact degree of each device and identifying the faulty device; and obtaining the final fault location based on the topology information of the faulty device through cross-cluster association decomposition, causal edge weight adjustment, and graph structure simplification. This method can accurately and quickly locate the source of a fault in complex network environments with severe noise interference and tight device coupling.

This invention discloses a feature selection method and system based on a causal graph model in federated learning. The method includes: S1. Initializing the candidate set and default weights; S2. Estimating and updating the weights through local bootstrapping stability; S3. Calculating growth statistics and masks, and aggregating and selecting the optimal features to add to the candidate set; S4. Evaluating the pruning statistics of the features in the candidate set and aggregating and deleting false positive features; S5. Iterating through S3 and S4 until no features satisfy the gating condition are found, and outputting the final Markov boundary candidate set. Through symbol mask consistency gating, only candidate features that are consistently positive across clients enter the competition set, thereby reducing the impact of spurious correlations under heterogeneous distribution and improving the cross-client consistency and interpretability of the selected features.

This disclosure provides a data-driven and scalable method to discover cause-and-effect relationships in data from natural systems that include sparse data sets. This technique can learn a causal graph from heterogenous data sources by combining embeddings from real data and embeddings from simulated data generated by process-based models. The causal graph is used for what-if analysis in out-of-distribution settings. One application is understanding the factors that affect soil carbon. A causal model created by these techniques can be used to discover cause-and-effect relationships that affect soil carbon. This model has applications such as forecasting soil carbon for a future time point to help inform farm practices. Farm practices, like tilling, may be modified in response to predictions provided by the model.

The application relates to an automatic attack tracing method, a terminal device and a storage medium, the method comprising the following steps: receiving logs of different sources and classifying and sorting the logs; after word-level splitting of log entries, obtaining tokenized log entries; constructing an abnormal log identification model, the model comprising an embedding extraction module and a classification module, all the tokenized log entries are taken as a training set, and the model is trained in a form of synchronous federated learning; when receiving a log to be traced, firstly tokenizing the log, then inputting the tokenized log entry into the trained model to obtain an abnormal log entry identification result; taking an event type corresponding to the abnormal log entry as a suspicious event; constructing an independent causal graph corresponding to each suspicious event, fusing the independent causal graphs of all the suspicious events, and reconstructing an attack story based on the fused causal graph. The application can detect attacks and track attack links, and efficiently reconstruct an attack story.

This application discloses a method, apparatus, and device for auditing vulnerability mechanisms, relating to the field of software security technology. The method includes: extracting vulnerability causal elements and sorting out causal relationships from multi-source evidence associated with vulnerabilities, which is linked to corresponding nodes in a causal graph through globally unique evidence citation identifiers; obtaining a causal graph with vulnerability causal elements as nodes and causal dependencies as edges; extracting observed variables from the multi-source evidence based on the causal graph and constructing a structural equation model based on causal dependencies; solving the structural equation model to obtain the quantitative contribution of each vulnerability causal element to the occurrence of the vulnerability; performing a bidirectional mapping match between the vulnerability causal elements and a pre-set security compliance clause library based on the quantitative contribution to obtain a bidirectional mapping relationship; and integrating multi-source evidence, quantitative contribution, and bidirectional mapping relationship to obtain a structured audit report. This enables complete tracing and accountability for the explanation of vulnerability mechanisms.

This invention relates to a multivariate time series anomaly root cause localization method based on causal structure learning and simulated perturbation verification optimization, belonging to the field of data science technology. The method includes: preprocessing the original normal data, dividing it into a training subset and a validation subset; calculating the mean and standard deviation based on the training subset; normalizing both the training and validation subsets; performing offline causal modeling on the normalized training subset to construct a global directed causal graph with edge weights and time delay information; constructing simulated perturbation samples on the normalized validation subset, and combining them with normal validation samples without injected perturbations to construct a joint optimization objective function; obtaining the optimal parameter combination through parameter search optimization; and using the global directed causal graph and the optimal parameter combination to locate the anomaly root cause of the test samples and outputting the results. This invention can provide more complete root cause localization results in complex anomaly scenarios and improve the analysis capability for multi-source anomalies.

The application discloses a depression recognition method based on a multi-view capsule graph network, belongs to the field of electroencephalogram analysis in the field of biological medicine, and comprises the following steps: preprocessing an electroencephalogram signal; extracting node features; constructing a topological graph, a functional graph and a causal graph according to the node features; adaptively fusing the topological graph, the functional graph and the causal graph to obtain a fused graph; a bidirectional adjustment mechanism of a capsule graph network; constructing a loss function and a total training target of the capsule graph network; and a depression recognition process based on a dynamic routing mechanism of the capsule graph network.The application introduces the capsule graph network to structurally model topological structures and connection modes among electroencephalogram channels; the mutual relationship among the electroencephalogram channels is enhanced through adaptive fusion, so that the capsule graph network can more accurately capture the interaction and coupling features of brain regions; and the bidirectional adjustment mechanism is introduced to flexibly adjust the contribution degree of features, so as to enhance the information transmission capacity of the capsule graph network in a global context and the sensitivity and robustness of the capsule graph network to depression-related features.

The application provides an artificial intelligence-based financing guarantee anti-fraud method and system, relates to the technical field of artificial intelligence, and comprises the following steps: constructing a feature vector by acquiring enterprise multi-source heterogeneous data, establishing a bidirectional causal graph structure to represent the risk transmission relationship between enterprises, acquiring a risk transmission adjacency matrix by using a NOTEARS algorithm, constructing a risk distribution metric space by using a Wasserstein distance to perform distribution robustness optimization, training to obtain an enterprise risk representation vector, and combining an ensemble learning framework to evaluate fraud risk. The application can effectively identify fraud in a complex network and improve the accuracy and reliability of financing guarantee risk assessment.

This invention discloses a river and lake eutrophication status monitoring system, belonging to the field of intelligent water environment monitoring technology, including data acquisition, feature fusion, state inference, probability calculation, and map construction modules. The data acquisition module acquires multi-source heterogeneous raw sensing data of the monitored water area, forming multi-dimensional water body feature parameters. The feature fusion module inputs the parameters into a deep feature fusion network, outputting a high-dimensional fused feature vector through adaptive weighting and nonlinear mapping. An inference engine embedded with historical cases trains a causal graph model, activating model nodes and paths based on high-dimensional features, calculating the simulated concentrations of total phosphorus and total nitrogen, and the posterior probability distribution of the comprehensive trophic state index, ultimately constructing a spatiotemporal evolution map of eutrophication. This system can fully explore the inherent correlations of water body characteristics, accurately infer index changes based on probabilistic causal logic, and completely restore the spatiotemporal evolution of eutrophication in water areas, making the monitoring results more consistent with the actual water body state.

This invention provides an intelligent online anomaly detection method and system for gas turbines. The method includes acquiring historical operating data and real-time multi-source sensor data for typical operating conditions throughout the entire lifecycle; constructing a refined semantic space and physical causal graph for operating conditions; identifying operating conditions based on a lightweight classifier; and modeling causal relationships between components through dynamic routing of the causal graph. It dynamically calibrates heterogeneous features using a multi-head attention mechanism and residual gating units, and performs graph neural network inference along directed causal paths. Anomaly identification and localization are performed using a combination of data-driven and knowledge-guided approaches, and a consistency constraint mechanism is used to fuse decisions, ensuring high-confidence output. The anomaly attribution stage employs gradient-weighted activation mapping and causal counterfactual reasoning, ultimately generating a structured anomaly cause explanation text. This invention possesses high real-time performance, strong causal explanation, and closed-loop optimization capabilities, effectively supporting the intelligent fault diagnosis process for the operation and maintenance of gas turbines in power plants.

This invention relates to the field of power monitoring, specifically to a method and system for real-time analysis of edge computing power metering data. The method includes the following steps: collecting raw sampling data from all metering nodes in a distribution area, completing hardware-level clock synchronization and lossless spatiotemporal alignment; dividing causal-related node groups according to property topology, extracting and filtering low-dimensional core causal features; embedding rigid constraints in power physics, constructing a lightweight baseline causal graph, and calibrating the baseline effect thresholds for endogenous normal causal paths and exogenous abnormal causal paths; collecting real-time incremental metering data, completing standardized preprocessing and feature extraction; performing lightweight causal inference based on the baseline causal graph, calculating real-time causal effect values; and distinguishing between normal bidirectional power flow and abnormal events through two-level threshold verification, completing the classification and determination of electricity theft or metering faults. This invention effectively solves the pain points of traditional analysis, such as inaccurate data, insufficient computing power, ambiguous judgments, and lack of source tracing.

The application discloses an online service system abnormality detection and root cause analysis method and system based on causal inference and business load, belongs to the technical field of micro-service fault detection and root cause positioning, and has the technical scheme that the method collects distributed tracking data under normal flow of the online service system, constructs a causal graph facing the business load by using service call topology prior constraints, regresses the structural equation between each load modeling index based on the causal graph and estimates the residual error distribution benchmark, then receives the distributed tracking data in real time, calculates the structural residual of each load modeling index through the structural equation, further standardizes the residual error, aggregates the standardized residual error, judges the abnormality based on the residual error distribution benchmark, and positions the abnormal root cause node from the causal graph by using the root cause preference score after the abnormality occurs; the system is used for realizing the method; the structural equation between each load modeling index based on the causal graph is regressed, the residual error distribution benchmark is estimated, the precise detection and root cause positioning of the silent business fault are realized.

This application provides a method, apparatus, device, and medium for anomaly diagnosis and handling of large models, relating to the field of artificial intelligence technology. After a large model enters a designated processing stage of the current task, when a preset anomaly event is determined based on monitoring data, current anomaly description information is extracted from the monitoring data. Based on a pre-constructed causal graph, a target observation node corresponding to the current anomaly description information is determined. From the root cause nodes connected to the target observation node, a target root cause node is determined. The target root cause node can characterize the anomaly cause corresponding to the current anomaly description information, achieving automatic and accurate diagnosis of the anomaly cause. From a pre-constructed anomaly handling strategy library, a target handling strategy corresponding to the target root cause node is determined and executed, forming a complete anomaly self-healing closed loop. This reduces manual intervention, improves the efficiency of anomaly diagnosis and handling, increases the probability of successful large model task startup, and saves computational resources.

The application relates to a cloud-native component system operation configuration detection method and device, computer equipment and a computer readable storage medium. The method comprises the following steps: constructing attribute vectors of target components in each cloud-native system according to node attributes of the target components, and maintaining multivariate relationship edges between the target components; constructing an attribute graph of the cloud-native system according to the target components, the attribute vectors of the target components and the multivariate relationship edges; analyzing time series data corresponding to node attributes of the target components based on a causal discovery algorithm, determining causal edges between the target components, and constructing a causal graph corresponding to the cloud-native system; fusing and aligning the attribute graph and the causal graph to obtain an enhanced causal attribute graph of the cloud-native system; determining an abnormal problem required to be detected in the cloud-native system, tracing the abnormal problem in the enhanced causal attribute graph, and obtaining a root cause result. The method can improve the optimization efficiency and resource utilization rate of the cloud-native system.

This invention discloses an intelligent detection method for chain-penetrating faults based on AI algorithms, belonging to the field of intelligent fault diagnosis technology. The method includes: employing a multi-scale temporal neural network computation model to segment, aggregate, and encode events in a structured multi-rate temporal data stream, generating a multi-scale abnormal event set; constructing a hierarchical dynamic causal graph based on the multi-scale abnormal event set and applying scale-causal consistency constraints to generate a multi-time-scale fused causal graph; and using a reduced-order physical simulation method to perform counterfactual intervention and experimental alignment on the root causes at the ends of each path, obtaining a simulation-experimental deviation sequence for each chain-penetrating path. This invention further utilizes semantically aware reverse traversal and scoring pruning in the fused causal graph to select a set of candidate chain-penetrating paths that conform to the fault semantic logic and causal intensity distribution from a massive pool of potential causal paths.

This application provides an automatic verification method and system for ship entry and exit declaration information based on artificial intelligence, belonging to the field of artificial intelligence technology. This application acquires the entry and exit declaration information of the target ship and a historical declaration verification dataset. Based on the historical data, it constructs a historical declaration sequence and a historical declaration deviation sequence. The historical declaration sequence is input into a trained prediction model to obtain a predicted declaration habit vector. A causal discovery algorithm is used to analyze the historical declaration deviation sequence, constructing a declaration deviation causal graph where nodes represent declaration item deviations and edges represent causal relationships. The difference between the current declaration and the predicted habit is calculated to obtain the declaration deviation vector, which is then input into the causal graph as observation evidence. Target nodes are determined through causal inference. A deviation threshold is calculated based on the historical deviation sequence. The current deviation value of the target node is compared with this threshold to determine the corresponding declaration information as abnormal declaration information, effectively identifying systematic and correlated declaration deviations in entry and exit declaration information.

The application relates to the field of intelligent transportation, and discloses an expressway intelligent monitoring and management system, method and electronic equipment. The system collects expressway global space-time data to construct a digital twin body synchronized with the physical world. In the twin body, potential risks are identified through prediction and deduction based on a space-time causal graph. In response to the risks, an optimal intervention strategy is generated through counterfactual deduction and executed, and the prediction intervention effect is recorded. After the intervention, the real traffic state and the prediction effect are compared, the counterfactual error is calculated, and the space-time causal graph is dynamically self-corrected accordingly. The application integrates perception, prediction, decision-making, execution and feedback into an adaptive control loop, introduces a self-correction mechanism based on counterfactual error, enables the system to continuously learn and evolve from interaction with the physical world, and solves the problems of model solidification and poor adaptability of traditional traffic management systems.

The application relates to the technical field of deep learning, in particular to a multi-user electricity consumption prediction method and model based on causal guidance. By using a causal discovery and multi-hop enhancement mechanism, the interference of pseudo-correlation factors between variables in a source data set is removed; in a feature propagation stage, static causal constraints and dynamic data-driven correlations are fused, and noise reduction propagation is performed under real causal path constraints, so that random fluctuations and noise interference caused by cross users are inhibited; in a prediction correction stage, a multi-hop spatial correction is performed on the prediction result based on a causal graph convolution correction unit, and cumulative prediction error divergence in a long sequence model is avoided. The application aims to solve the problems of how to avoid pseudo-correlation interference caused by shared exogenous influencing factors in multi-variable time series prediction and error accumulation leading to prediction accuracy degradation in long sequence propagation.