139 results about "Interactive analysis" patented technology

The invention relates to the field of government affair work order intelligent processing, in particular to a government affair work order intelligent processing method and system based on space-time semantic clustering and a large language model. According to the scheme, unified data feature modeling is conducted on a work order to be processed, an improved DBSCAN clustering algorithm is executed on the work order through a weighted space-time semantic three-dimensional distance measurement formula, and combined clustering of space, time and semantic features is achieved; calculating priority scores of the work orders, and dynamically allocating scheduling resources according to the clustering scale and the priority of the work orders; based on a retrieval enhancement generation technology of an RAG framework and an FAISS vector retrieval library, historical similar work orders are matched, a few-sample learning case is generated, and two sets of differential treatment schemes are generated by controlling temperature parameters of a large language model; visual display and interactive analysis of work order clustering are realized through an interactive GIS platform; and establishing a quality feedback closed loop of work order reconstruction. The method is suitable for intelligent government affair work order processing.

Data platforms described herein are configured to monitor a compute environment and facilitate interactive analysis of multifaceted security threats within the compute environment. Such a data platform may determine that one or more assets within the compute environment are possibly being targeted by a multifaceted security threat and present an interactive user interface. The user interface may be configured to display an identifier indicative of the multifaceted security threat, a set of selectable evidence items each associated with a different facet of the multifaceted security threat and assessed based on the monitoring of the compute environment, and a presentation pane for displaying information. As such, the data platform may detect a selection of a particular evidence item from the set of selectable evidence items and, in response to the selection, populate the presentation pane with information related to the particular evidence item. Corresponding methods, systems, and products are also disclosed.

The invention discloses a retrieval enhancement question and answer generation method and system based on a large language model in combination with a multi-agent cooperation mechanism. The method comprises the steps of S1, text preprocessing and semantic difference enhancement and amplification; s2, checking and complementing user questions; s3, evaluating problem complexity and selecting a generator; s4, text retrieval and answer generation; the system comprises an information enhancement processing module, a query understanding optimization module, a question complexity intelligent evaluation module and a collaborative answer generation module, and is used for realizing the method. According to the method, a multi-agent collaborative RAG framework of an information enhancement agent, an interaction analysis agent and a problem complexity evaluation agent is introduced, and semantic difference enhancement and amplification, query integrity verification and complementation and a dynamic generation strategy based on problem complexity are performed after hierarchical document partitioning are combined; and the performance of the system in the aspects of semantic distinguishing capability, query understanding precision, generation efficiency and reliability is comprehensively improved.

The invention discloses a sensor-based GIL pipe gallery expansion joint on-line monitoring system and method, and relates to the technical field of power equipment monitoring, and the method comprises the steps: collecting key state information and environmental parameters of an expansion joint through the deployment of an expansion joint sensor and an environmental sensor, and carrying out the preprocessing; performing deep fusion on expansion joint sensor data and environment sensor data, constructing a comprehensive state evaluation and prediction model, and analyzing the influence of different factors including environment, mechanical faults and material aging on the expansion joint state; correcting state parameters for environmental factors; for mechanical faults, fault types, positions and severity are identified, and corresponding early warning is triggered; aiming at material aging, monitoring and predicting performance degradation and residual life in real time, and formulating a maintenance strategy; the model performance is evaluated and updated regularly, an early warning threshold value is adjusted dynamically, and self-adaptive early warning is achieved; a data integration and visualization platform is developed, and visual display and interactive analysis of monitoring data are achieved.

The embodiment of the invention provides a network spoofing defense method and system based on AI virtual topology generation, and the method comprises the steps: continuously collecting multi-source heterogeneous data in a network, and constructing and dynamically updating a real network topology portrait based on the multi-source heterogeneous data; on the basis of a generative adversarial network model, adversarial training is carried out by taking the real network topology portrait as a training sample, the structure and characteristics of a real network are learned, and a dynamically changing simulation virtual network topology is generated; seamless fusion and dynamic deployment are carried out on the simulation virtual network topology and the real network, and a trapping environment is constructed; monitoring and recording an attack behavior of an attacker in the trapping environment, and performing deep interaction analysis on the attack behavior and generating threat intelligence; and based on the threat intelligence, dynamically optimizing a deception strategy and executing a closed-loop control response. According to the embodiment of the invention, the virtual topology which is highly similar to a real network but is completely imaginary can be dynamically generated by utilizing artificial intelligence.

The invention relates to the technical field of natural language processing, and discloses a special disease knowledge spectrogram construction and analysis system, which comprises a federal data acquisition module, a dynamic ontology mapping preprocessing module, a cross-source knowledge collaborative extraction module, a self-adaptive graph construction module, an intelligent dynamic updating module and an interactive analysis application module, the cross-source knowledge collaborative extraction module is used for collaborating multi-source data on the premise of guaranteeing data privacy, the dynamic ontology mapping preprocessing module is used for mapping knowledge representation of the multi-source data to a unified standard, and the cross-source knowledge collaborative extraction module is used for extracting medical entities and semantic relationships based on multi-modal data fusion. According to the method, through the cooperation of the longitudinal federated learning framework and the secure multi-party computing algorithm, the problem of multi-source medical data privacy collaborative computing is solved, and the advantages of safely aggregating multi-mechanism data and providing original data support for map construction on the premise of ensuring data privacy are achieved.

The invention discloses a concrete foundation pile visual information integration system and method based on 3D GIS and ultrasonic tomography fusion, the 3D GIS and ultrasonic tomography technology are fused, internal defects of a foundation pile are accurately positioned and visually presented, and the detection intuition and accuracy are remarkably improved. By constructing a three-dimensional information model, correlation analysis of foundation pile defects, structures and geologic body information is achieved, and engineers are assisted in comprehensively evaluating structural safety and risk levels. The system adopts an optimization algorithm and a multi-level framework, not only improves data processing efficiency, but also supports multi-dimensional interaction analysis, provides a scientific basis for operation and maintenance management and reinforcement design of the foundation pile, and effectively overcomes the defects that a traditional detection means is difficult to position, depends on experience and lacks unified management.

The invention belongs to the technical field of underground engineering and computer-aided engineering analysis, and particularly relates to an anti-seismic safety collaborative analysis method for multi-modal data of proximity tunnel engineering. By integrating engineering geology, structural design, construction monitoring and multi-level seismic oscillation data, a three-dimensional numerical model fusing geology and structural information is constructed. Firstly, refined simulation of the whole construction process and the operation static load is carried out, and a reference mechanical state is obtained; carrying out multi-task condition earthquake dynamic time history analysis on the basis, dynamically capturing dynamic response of surrounding rock and a structure, carrying out collaborative interpretation on static and dynamic analysis results, emphatically analyzing spatio-temporal evolution and connectivity risks of a tunnel anchor and an adjacent highway tunnel surrounding rock plastic zone, and extracting a plurality of quantitative collaborative indexes; and an interactive analysis report integrating a three-dimensional visual scene, a multi-dimensional comparison chart and a rule base-based safety interpretation conclusion is automatically generated, and systematic decision support is provided for seismic design, construction control and safety evaluation of proximity projects.

The invention discloses a software supply chain vulnerability assessment method based on multi-Agent collaboration and dynamic hypergraph learning, which is used for opening a link between vulnerability assessment and repair and realizing full-life-cycle collaboration linkage. For the problem of vulnerability assessment and repair splitting, the method covers the core capabilities of static analysis (SAST), dynamic detection (DAST), interactive analysis (IAST), intelligence analysis and the like. Evaluation results of various agents are output through standardized vulnerability portraits, and a subsequent repair strategy is driven to be automatically generated and executed, so that whole-process closed-loop linkage and fine control of'discovery-evaluation-repair-verification 'of vulnerabilities are realized.

The invention discloses a marine geological environment multi-dimensional data reconstruction and dynamic visualization mapping method and system, and belongs to the technical field of marine information technology and data visualization. In order to solve the problems that in the prior art, standard data asset structure is isolated, heterogeneous data deep fusion is difficult, and mass data visualization performance is bottleneck, the method mainly adopts the following technical scheme that standardized multi-source data is received, a data volume mapping rule is established, discrete observation data is reconstructed and fused into a unified and continuous three-dimensional data volume, and the three-dimensional data volume is obtained; and executing a three-dimensional space pattern recognition algorithm on the three-dimensional data volume to recognize a feature entity, and providing on-demand query and multi-resolution access services through a visual mapping interface. According to the method, efficient automatic modeling and high-performance interactive analysis of mass and multi-dimensional marine geological environment data can be realized.

The embodiment of the invention provides a server test platform, method, device and equipment and a storage medium, an interaction analysis system of the server test platform is used for receiving to-be-tested information input by a user, performing semantic analysis on the to-be-tested information to obtain a semantic analysis result, and sending the semantic analysis result to the server test platform; the semantic analysis result is used for determining a target test to be executed on the server test platform; displaying a test result of the target test in the interactive analysis system according to a target form through the target abnormal mode and the multi-dimensional dynamic association data; the three-dimensional association model is used for extracting multi-dimensional dynamic association data about the target test in the process of executing the target test; identifying a first abnormal mode appearing in the target test through the multi-dimensional dynamic association data; the prediction analysis model is used for predicting a second abnormal mode occurring in a target test corresponding to the multi-dimensional dynamic association data; the knowledge management system is used for determining a diagnosis scheme of a test result through the knowledge graph, so that the test efficiency and the test effect are improved.

The invention discloses an AI engine docking decision-making method based on a 5G new call, and the method comprises the steps: carrying out the call service, and analyzing the call information of a user; the AI engine provides an intelligent interaction analysis capability; the media docking module is used for receiving a function instruction of calling service logic, providing basic media capability and making and displaying media content according to user operation; the BM module is used for storing user services and setting data; wherein the call service synchronizes a call instruction, audio and video media stream information and user behavior data to the media docking module, and receives a media instruction. According to the method, real-time multi-mode interactive perception and dynamic AI engine docking can be realized, so that the bottleneck in the prior art is broken through.

The invention relates to methods and systems for integrating generative artificial intelligence (AI) capabilities into Software as a Service (SaaS) platforms. It comprises maintaining AI agents with varying credentials, enabling their interaction with alphanumeric data in table structures, and implementing a hierarchical access control scheme. The system displays table structures, provides interfaces for user inputs, and allows AI agents to be added as platform users. The generative AI agents can analyze data, identify actions, and perform tasks autonomously. The invention also includes methods for proactive information gathering, interactive analysis of AI outputs, and management of AI resources as limited assets. This approach enhances SaaS functionality by enabling AI-driven task completion, data analysis, and decision-making while maintaining data security and user-specific access controls.

The invention discloses a multi-dimensional data association analysis system and method based on unified modeling, and relates to the technical field of big data intelligent analysis. Comprising the following steps: 1, constructing a semantic model; step 2, carrying out query compiling and optimization; 3, performing query execution and scheduling; 4, interactive analysis and visualization are carried out, wherein a multi-dimensional report designer MRD is used for providing end-to-end full-process visualization display, and interactive drilling and narrative analysis are supported; step 5, performing closed-loop optimization; according to the method, the complexity of an underlying data structure is shielded through the semantic model, and automatic correlation analysis of a cross-fact table is supported, so that business personnel can independently complete deep data exploration and decision support without mastering professional technologies such as SQL (Structured Query Language) and the like.

The invention provides an intelligent face document personalized recommendation method and system. The method comprises the following steps: firstly, collecting multi-dimensional face document demand data (including historical consumption, real-time demand and preference change data) and face document supply data (including real-time supply, feature change and consumption feedback data) of a user; constructing a demand evolution trajectory based on the user demand data, and presenting a demand change trend and an association relationship; constructing a supply feature time sequence association set based on the supply data, and presenting a supply feature association mode and a change rule; calling an artificial intelligence model for interactive analysis to generate a dynamic demand supply matching rule; according to the method, the rule is iteratively optimized by using the user feedback data, and the personalized recommendation result of the face file is finally generated, so that the accurate dynamic matching of the demand and the supply is realized, and the user satisfaction and the face file operation benefit are improved.

The invention discloses a construction method of a single-cell multi-omics data analysis system based on a containerization technology, and the construction method of the single-cell multi-omics data analysis system based on the containerization technology comprises the following steps: constructing a single-cell multi-omics analysis modular toolkit; a production mirror image is constructed on the high-performance cluster through a Single container engine, and deployment of a production environment is completed; a single-cell multi-omics automatic analysis platform is built through a snakemake process engine; and a JupyterLab interactive analysis platform for exploratory analysis of the single-cell multi-omics data is constructed. According to the single-cell multi-omics data analysis system construction method based on the containerization technology, the data acquisition link in the analysis process is accelerated, an automatic working process is formed to solve the problem of complex process, and the actual requirements of multiple users on a high-performance computing cluster for scientific exploration and analysis of single-cell data are met.

The invention provides a geological exploration data visualization dynamic processing method and system, and relates to the technical field of intelligent geological exploration, and the method comprises the steps: constructing a prior three-dimensional geological structure model and a forward substitution model; constructing a multi-source joint loss function based on forward modeling output of the forward modeling substitution model and geological survey data; based on a multi-source joint loss function, performing layer-by-layer reasoning reconstruction on geological point cloud parameters in the prior three-dimensional geological structure model to obtain a dynamic three-dimensional geological parameter set; performing fusion rendering and interaction analysis on the dynamic three-dimensional geological parameter set to obtain a visualization result of a fusion geological structure and multi-field evolution information, and constructing a scene deduction component based on the dynamic three-dimensional geological parameter set and a forward substitution model; according to the geological situation sensing and deduction early warning method and device, high-precision geological situation sensing and deduction early warning can be achieved through deep fusion of the forward modeling model constrained by physical information and the layered dynamic inversion algorithm.

The invention relates to the technical field of virtual reality application, in particular to a virtual reality assisted urban and rural spatial layout visualization planning system, which comprises a data acquisition and preprocessing module, a three-dimensional modeling module, a virtual reality visualization module, a planning analysis module and an interaction and feedback module. Wherein the data acquisition and preprocessing module is used for acquiring and preprocessing urban and rural spatial data; the three-dimensional modeling module is used for constructing a three-dimensional urban and rural model based on the preprocessed spatial data; the virtual reality visualization module is used for importing the three-dimensional urban and rural model into a virtual reality environment to generate a VR visualization scene; the planning analysis module is used for performing spatial layout analysis and outputting a planning analysis result; and the interaction and feedback module is used for receiving user interaction input and updating the planning scheme in real time. According to the urban and rural spatial layout planning method and system, by integrating the processes of data preprocessing, three-dimensional modeling, virtual reality visualization and dynamic interactive analysis, visual expression and real-time optimization of urban and rural spatial layout planning are achieved.

The invention relates to a large sample exploratory experiment conclusion macro-micro interaction verification method and system, the method is used for constructing a'macro-interaction-micro 'three-layer linkage analysis framework, and the framework comprises a macro conclusion layer, an interaction analysis layer and a micro instance layer; the interaction analysis layer is used for connecting the macroscopic conclusion layer and the microscopic instance layer; the method comprises the four steps of automatic discovery and presentation of a macroscopic conclusion, interactive deep exploration and verification based on the macroscopic conclusion, macro-micro bidirectional linkage and sample screening, and concrete redisk and causal verification of micro instances. Through the method, interactive verification of a macroscopic conclusion and a microcosmic instance is realized, and the interpretability and credibility of the conclusion are remarkably improved. According to the method, a macroscopic and microcosmic barrier is broken through, a cognitive closed loop from'knowing the course 'to'knowing the course' is constructed, and a more robust basis is provided for decision making.

The invention discloses a spatial intelligent scene early warning method and system based on three-layer cooperative constraint, relates to the field of spatial modeling, and solves the problem that the spatial intelligent scene early warning method is poor in early warning effect. S2, establishing a scene three-dimensional coordinate system and a model three-dimensional coordinate system according to an acquisition result, performing interactive analysis on the scene three-dimensional coordinate system and the model three-dimensional coordinate system, and obtaining regional laser point division data according to an analysis result; s3, performing scene intelligent analysis on the target early-warning area according to the early-warning area real-time interaction model, issuing space intelligent early warning according to an analysis result, and establishing an early-warning area real-time interaction model on the basis of the model edge sampling points. According to the invention, the accuracy of a space intelligent early warning result can be improved.

This application discloses a method, device, and electronic device for displaying energy flow in a power plant, belonging to the field of new energy power plant technology. The method includes: generating an energy flow diagram corresponding to the current power plant configuration and displaying it on an interface; the energy flow diagram includes multiple visual elements, each visual element including at least source nodes and carrier nodes representing power plant equipment, and path elements representing the energy flow relationship between source nodes and carrier nodes; responding to user interaction with any node or path element, determining the energy path where the interacted node or path element is located; updating the rendering state of the energy path on the interface to make the energy path appear locally focused relative to other energy paths. This application achieves energy path-level visualization and interactive analysis based on power plant equipment topology, improving the accuracy of expressing power plant energy flow relationships and the efficiency of analyzing multi-dimensional operational data.

The invention relates to the technical field of high-current particle beam diagnosis, in particular to a high-energy ion beam three-dimensional reconstruction method and system based on secondary electron emission signals. According to the technical scheme, the method comprises the steps of data preprocessing and conversion, geometric space up-sampling, point assignment generation and model boundary division and output. According to the method, the spatial sparsity of original data is overcome through geometric upsampling, the fidelity of physical attribute reconstruction is remarkably improved by utilizing a topological structure based on a minimum spanning tree and a primary function frequency superposition model, automatic and accurate division of beam boundaries is realized according to a physical threshold value, an end-to-end automatic processing flow is formed, and the method has the advantages that the method is simple and convenient to operate and high in practicability. And outputting a visual three-dimensional model supporting interactive analysis, systematically solving the problems of data sparsity, interpolation distortion and boundary fuzziness in the prior art, and realizing high-precision and high-efficiency reconstruction and visual analysis of the three-dimensional form of the high-energy ion beam.

The application relates to the technical field of data security processing, and discloses a safe interaction analysis processing system for cross-department supervision collaborative data, which comprises a strategy fingerprint generation module, an access authorization issuing module, a data compliance packaging module, a security verification loading module, a joint accounting analysis module and a result output auditing module.The strategy fingerprint generation module is used for generating a strategy fingerprint based on task information; the access authorization issuing module is used for performing attribute authorization and issuing a bill according to the strategy fingerprint and a user identity; the data compliance packaging module is used for generating a compliance view according to the strategy fingerprint and performing deterministic desensitization and encryption packaging; the security verification loading module is used for verifying the consistency of a packaging bag and loading the packaging bag into a temporary session library to generate a session fingerprint; the joint accounting analysis module is used for performing deterministic joint accounting and early warning determination; and the result output auditing module is used for clipping an accounting result and generating an auditable proof bag.The early warning distribution monitoring module is used for pushing early warning information and monitoring a task execution state.The application realizes trusted interaction, compliance analysis and traceable management of cross-department supervision data under the condition of scattered storage.

The invention discloses an enterprise financial risk real-time early warning and visualization method based on stream-oriented computing, and the method comprises the following steps: 1) multi-source heterogeneous data real-time collection and access: collecting stream-oriented data from a plurality of data sources in real time, the data sources comprising internal business data and external financial data; 2) stream data cleaning and standardization: preprocessing the data accessed in real time to ensure that the formats of the data from different sources are uniform and can be used for subsequent calculation; 3) real-time financial index calculation and risk model evaluation: 3.1) real-time index calculation; (3.2) carrying out real-time risk assessment; 4) risk early warning and graded pushing; 5) performing real-time visualization and interactive analysis; according to the method, efficient analysis, visual presentation and man-machine interaction of complex data are achieved in an automatic and intelligent mode, and therefore support is provided for decision making.

The application discloses a multi-peak decomposition time-resolved Raman spectrum thermal map interactive analysis method and system, and belongs to the field of spectrum analysis. The method comprises the following steps: continuously or time-divisionally collecting a to-be-measured sample to obtain Raman spectrum data corresponding to different time points; preprocessing the original Raman spectrum data; extracting spectral characteristic peaks by using an adaptive peak value recognition algorithm, and performing characteristic peak integral calculation; obtaining time contribution coefficients of different chemical components or characteristic signals in the time dimension by using a multi-peak decomposition algorithm; generating a Raman thermal map in the time dimension according to the characteristic peak intensity and / or the time contribution coefficients; establishing an index mapping relationship between the position of the Raman thermal map and the corresponding time Raman spectrum data; and responding to the interactive operation of the user on the thermal map and displaying the Raman spectrum curve at the corresponding time point in real time. The application can realize visual display, dynamic monitoring and interactive analysis of time sequence data of Raman spectrum, and improves the efficiency and accuracy of Raman signal analysis of dynamic samples.

The invention relates to the technical field of power grid project cost management and business intelligence, and discloses a project cost multi-stage intelligent analysis and examination method, system, equipment and medium, and the method comprises the steps: building a power grid project cost data warehouse, and achieving the integration and standardization of multi-source cost data; establishing an OLAP multi-dimensional analysis model, and supporting general cost comparison and key index analysis in a budget calculation stage, feasibility research estimation, preliminary design budget calculation, completion settlement three-calculation comparison, first supply material check and review list consistency check in a settlement stage; designing an intelligent decision engine based on data mining, and generating review opinions through association rules, a classification algorithm and an anomaly detection technology; and a BI visual instrument panel is developed, and multi-dimensional display, real-time monitoring and interactive analysis of the power grid cost data are realized. Through systematic application of the BI technology, deep mining, intelligent decision and visual display of engineering cost analysis are realized, and the efficiency and precision of cost review are remarkably improved.

The application discloses a code automatic generation method for power grid dynamic intelligent calculation and interactive analysis, comprising the following steps: formulating a code generation plan for power grid calculation problems; generating a real-time state estimation result of power grid dynamic topology embedding according to real-time power grid measurement data; analyzing the structured instructions of each subtask in the code generation plan, fusing the structured instructions with the power grid dynamic topology embedding to construct an enhanced Prompt to guide a double-teacher model of a large language model and a simulation mechanism model to generate a <Q, think, A> data set of code sequences, taking a base large language model as a student model, selecting samples from the data set to train the student model, then selecting a preferred data set from the data set to conduct reinforcement learning on the student model in combination with the power grid dynamic topology embedding, and generating codes according to the code generation plan and the power grid dynamic topology embedding by using the trained student model. The application significantly improves the automation and intelligent level of power grid calculation.

The invention relates to the technical field of meteorological data processing and visualization, aims to solve the problems of low format conversion efficiency, rigid fusion strategy, cross-source collaborative fault and insufficient interaction in existing multi-source meteorological cloud cover forecast data processing, and provides a fusion processing and interaction analysis method. The fusion processing method comprises the following steps of: receiving a GFS and ICON double-source compressed package (0.25-degree resolution, one file per hour and 7-day data), decompressing and backing up in parallel, directly analyzing a binary file to generate a single-source cloud picture, fusing according to a configurable proportion alpha (0 < = alpha < = 1, default 0.5, and alpha automatically complemented to 1 or 0 when a certain source is missing), generating a visual cloud picture, and storing the visual cloud picture in a corresponding database. The interactive analysis method comprises the following steps: receiving a user space, time and data source / alpha three-dimensional query request, generating a 168-hour three-source comparison curve when point query is less than or equal to 1 second, displaying a fusion cloud picture and a double-source transparent adjustment picture by region query, and supporting single-interface synchronous comparison. The method improves the processing efficiency and scene adaptability, and improves the decision-making efficiency.

The invention relates to the technical field of modern electric power systems, and particularly discloses an electric power high-voltage data monitoring and waveform interactive analysis system, which comprises a data receiving layer, an event driving layer, a business processing layer and a front-end rendering layer, the data receiving layer is used for receiving the power high-voltage waveform data in real time and transmitting the power high-voltage waveform data to the event driving layer; the event driving layer is used for dispatching the electric power high-voltage waveform data and transmitting the dispatched electric power high-voltage waveform data to the business processing layer; the business processing layer is used for processing the scheduled power high-voltage waveform data and transmitting the processed power high-voltage waveform data to the front-end rendering layer and the time sequence database; and the front-end rendering layer is used for carrying out rendering optimization on the processed electric power high-voltage waveform data and returning a touch instruction to the business processing layer. According to the invention, the accuracy and real-time performance of power high-voltage monitoring can be greatly improved, and a convenient and powerful waveform analysis tool can be provided for power engineers.

The invention discloses an intelligent legal file retrieval method and system based on a multi-modal heterogeneous large model, and the method comprises the steps: obtaining original legal file data containing a text, an image and a table, and carrying out the multi-modal analysis and structured preprocessing; extracting depth features of each modal through a multi-modal heterogeneous large model, and fusing the depth features by adopting a multi-modal cross attention mechanism; constructing a case element sparse graph and optimizing the case element sparse graph through an optimization algorithm; carrying out weight adjustment and vector optimization by using a robust learning real-value multi-index model algorithm; candidate case subsets are screened through a random recovery sampling technology, and multi-dimensional matching calculation is carried out; and carrying out visual presentation and interactive analysis processing on the similar historical case list. Through the multi-modal fusion understanding, sparse graph optimization, robust learning and other technologies, the legal file intelligent retrieval accuracy and efficiency are effectively improved, and accurate case reference and decision support are provided for legal workers.