788 results about "Relationship analysis" patented technology

The invention belongs to the technical field of bridge diagnosis, and discloses a bridge crack intelligent diagnosis system based on multi-modal data fusion. By fusing multi-source data such as visual images, sound wave detection and vibration signals, comprehensive perception and characterization of crack features are realized; constructing a bridge crack characteristic spectrum diagram by adopting a cross-modal feature extraction and heterogeneous feature coding technology; generating a crack evolution situation map based on space-time correlation analysis and knowledge graph construction; the robustness of the system in a complex environment is improved through environmental adaptability feature enhancement and multi-scale characterization; constructing a bridge safety risk hypergraph in combination with multi-dimensional risk analysis and multi-agent collaborative diagnosis; analyzing and revealing a crack evolution mechanism by applying a causal relationship; and finally, through dynamic fusion and uncertainty quantification, a crack intelligent diagnosis comprehensive report is generated. According to the system, the limitation of traditional single-mode diagnosis is broken through, and dynamic prediction and accurate risk assessment of fracture evolution are realized.

The invention relates to a power equipment fault prediction method based on multi-source heterogeneous data and a twinborn model, and the method comprises the following steps: S1, obtaining multi-modal data related to the operation of power equipment, constructing a joint feature encoder through an attention mechanism in deep learning, and carrying out the joint modeling of multi-modal information, and obtaining a joint feature representation; s2, based on joint feature representation, adopting a dynamic twin model of self-supervised learning and fusion reinforcement learning, and dynamically updating the equipment health state; s3, constructing a complete-cycle prediction framework based on Transform, capturing long-term dependency and dynamic change characteristics in equipment operation, and identifying trigger factors of faults in combination with causal relationship analysis; and S4, based on the equipment health state and the fault prediction result, optimizing equipment operation parameters by adopting an intelligent operation and maintenance optimization engine, and dynamically generating an optimal operation and maintenance plan in combination with a genetic algorithm. The reliability and the operation efficiency of the power equipment are remarkably improved.

The invention discloses a pharmaceutical quality traceability decision-making method and system based on a dynamic knowledge graph, and relates to the technical field of pharmaceutical quality control. The method comprises the following steps: defining a dynamic knowledge graph containing quality-related entities, wherein the dynamic knowledge graph at least comprises a quality event entity, a production element entity and a time sequence causal relationship between the quality event entity and the production element entity; capturing production data related to quality in real time, and generating event data with a quality rule identifier based on a preset triggering condition; and carrying out dynamic state modeling on the production element entity to generate a state vector containing a time sequence attenuation characteristic and a quality influence weight. By constructing the dynamic knowledge graph and designing the time sequence causal reasoning algorithm, real-time capture, dynamic modeling and causal relationship analysis of production data are achieved, the defects of the traditional technology in real-time performance and relevance are overcome, and an efficient and accurate intelligent solution is provided for pharmaceutical quality traceability.

The invention relates to a membrane pool optimization control method, system and equipment based on multi-agent collaborative decision-making and a medium, and the method comprises the steps: abstracting each membrane pool into an agent, building a state space and an action space, and creating a global reward function; applying hard constraint and dynamic soft constraint to the action space to obtain a limited action space; an agent relation analysis layer is embedded in the evaluation network, behavior characteristics and cluster cooperation intensity distribution of all agents are obtained through high-dimensional characteristic mapping and a multi-head attention interaction mechanism, and global value evaluation is generated by combining rewards of a global reward function and through cooperation intensity distribution correction; and synchronously optimizing strategy parameters of each agent based on global value evaluation by adopting a double-stage collaborative training framework, and injecting controllable exploration noise to drive collaborative convergence of the membrane pool cluster. The purposes of saving energy, reducing emission, reducing cost, improving efficiency, reasonably distributing productivity and prolonging the average service life of the membrane module are achieved, and water plant operation is obviously promoted to develop towards the intelligent and sustainable direction.

The invention relates to the technical field of data monitoring, in particular to an automatic monitoring system for plasma meltallizing coating process data, which is characterized in that a parameter association relationship analysis module performs association analysis based on real-time parameters in a plasma meltallizing coating process, identifies an association relationship among the parameters and generates a process parameter association analysis result; according to the method, through multi-dimensional correlation analysis and path screening of the technological parameters, key factors influencing the coating quality and action links of the key factors are defined, and the abnormal fluctuation state and the transfer path can be accurately predicted in combination with parameter priority dynamic adjustment and real-time data feedback, so that the stability and continuity in the technological process are ensured. In the scheme, a dynamic weight distribution strategy is adopted, so that key parameters in an abnormal state occupy a higher priority position, parameter distribution is rebalanced when the normal state is recovered, and excessive interference of a single parameter on the whole process is avoided.

The invention relates to the technical field of thermal power production, artificial intelligence and time sequence analysis, in particular to a measuring point time sequence anomaly analysis method and system based on a multi-modal large model. Calculating a cross-modal feature similarity weight based on a self-attention mechanism to obtain a fusion feature vector; comparing the fusion feature vector with the normal mode feature representation through a self-supervised learning algorithm to realize anomaly detection, wherein a result comprises anomaly time, a fault image area and a text keyword; and a result is input into a multi-modal large model, cross-modal attention mechanism association data is utilized, fault reasoning and causal relationship analysis are carried out, a visual diagnosis report is generated, and then the accuracy of measuring point screening and time sequence anomaly analysis is improved.

The invention provides a production line process scheduling optimization method and system based on an agent cluster. The method comprises the steps that the agent cluster comprising a plurality of distributed decision agent nodes is configured according to a production line process scheduling instruction; based on the production line architecture feature graph, a communication protocol between agent nodes is generated through node association relationship analysis, and a cooperation rule set is constructed in combination with a process priority rule and a resource constraint condition; the intelligent agent cluster is instructed to perform scheduling according to a communication protocol and a cooperation rule set, process tasks are dynamically allocated and adjusted through real-time state interaction among nodes, scheduling behavior logs are obtained, a reinforcement learning training sample set is constructed in combination with expected labels, decision parameters of intelligent agent nodes are optimized by using a multi-intelligent-agent reinforcement learning algorithm, and a reinforcement learning training result is obtained. According to the method, the dynamic response capability and the cooperation efficiency of automatic production line scheduling can be improved, and the method is suitable for efficient process scheduling in a complex production environment.

The invention relates to the technical field of AGV scheduling control, and discloses an AGV scheduling control method and system for automobile carrying, and the method comprises the steps: carrying out the data collection of a plurality of AGVs in a stereo garage, and obtaining a state information data set; performing causal relationship analysis on each AGV motion track to obtain a dynamic space-time causal reasoning graph; performing trajectory feature extraction and operation state prediction through a time convolutional neural network to obtain an operation trajectory prediction sequence of each AGV; carrying out deadlock loop identification in a local adjacent AGV and global garage range to obtain a deadlock risk state information matrix; according to the multi-AGV cooperative scheduling optimization method, the multi-AGV vehicle carrying task allocation and the driving path selection are subjected to anti-deadlock optimization to obtain a multi-AGV cooperative scheduling optimization strategy, the efficient cooperative execution of the multi-AGV vehicle carrying task in the stereo garage is realized, and compared with the prior art, the conflict and waiting time between the AGVs can be obviously reduced, and the operation efficiency and the service quality of the stereo garage are improved.

The invention relates to the technical field of equipment monitoring and management, and particularly discloses an AI oT-based distributed industrial equipment monitoring and management system and method, and the system comprises a distributed equipment monitoring module which collects operation parameters, performance indexes and environment data in real time through a plurality of types of sensors disposed on industrial equipment, a dynamic hierarchical edge architecture is utilized to process data and evaluate a static health state, and task intelligent layering and cloud collaboration are supported. And the multi-modal perception analysis module generates an equipment association graph based on the collaborative task, extracts multi-modal features by using a cross-node collaborative algorithm, and realizes dynamic health assessment by combining a CNN-LSTM fusion model and equipment implicit dependency relationship analysis. And the distributed equipment management module fuses equipment dynamic and static health data, cooperates with a task association graph and a fault tracing model, and realizes efficient resource configuration and fault prevention. The system guarantees intelligent monitoring and management of the whole life cycle of industrial equipment through a cooperation mechanism of edge intelligence and cloud deep analysis.

The invention relates to the technical field of distributed agent collaboration and graph database query optimization, in particular to a graph query processing method and system based on multi-agent collaboration, and the method comprises the following steps: reasoning a coordination agent to analyze the semantics and intention of user query by using a large language model; a structured multi-agent execution plan is output, and an entity search agent or a relation analysis agent needing to be called is explicitly called; the agent coordinator receives the multi-agent execution plan and selects agents according to task types and task allocation; obtaining a distributed intermediate result set; starting a result aggregation mechanism, and performing typed processing, priority ranking and deduplication merging on the distributed intermediate result set to obtain an arrangement result; and outputting an answer for answering the initial question of the user. The system comprises an agent execution plan output module, a multi-task execution module and an arrangement result reasoning module. According to the method, efficient, extensible and fault-tolerant graph query processing is realized.

The invention discloses an outbound intelligent routing service system, relates to the technical field of routing service, and aims to solve the problem that data quality monitoring is incomplete in a calling process. According to the method, the calling queues are sequenced, and the outbound request is initiated according to the priority, so that the high-priority task can be preferentially processed. The call-out efficiency is improved, the overall service performance is improved, when a quality problem occurs in the call process, active adjustment is carried out by adopting a jitter buffering technology, the intelligent adjustment mode can quickly respond and improve the call quality, and when the quality problem is monitored, future call parameters of the client are automatically adjusted, so that the call-out efficiency of the client is improved. The adaptive adjustment strategy can ensure that the service quality of important customers is always kept at a relatively high level, the data analysis stage comprises descriptive analysis, causal relationship analysis and trend analysis, and the analysis methods can deeply mine information and rules in data from different angles.

The invention relates to the technical field of natural language processing, in particular to a hypergraph neural network-based multi-modal entity relationship extraction method, which comprises the following steps of: acquiring a text-image pair and inputting the text-image pair into a preset hypergraph construction model, acquiring node characteristics and a multi-modal hypergraph structure of the text-image pair, the hypergraph structure comprises a group of hyperedge sets, each hyperedge in the hyperedge sets can be connected with a plurality of nodes at the same time, and the number of the nodes connected with the hyperedges is not limited; the node features and the multi-modal hypergraph structure are input into a preset hypergraph neural network model, the multi-modal entity relation in the text-image pair is output, and the hypergraph neural network model is used for analyzing hypergraph node information propagation from the semantic and contextual relation and giving corresponding weights to semantic information between modals in combination with an attention mechanism. According to the method, the multi-modal information can be effectively fused, and the entity category and the entity relationship of the multi-modal information can be identified.

The invention relates to the technical field of water level and water surface relation analysis, in particular to a lake water level and water surface relation analysis method based on a DEM and a remote sensing image, which comprises the following steps of: based on DEM and remote sensing image data, analyzing a space relation between measured data and the DEM, correcting abnormal elevation and adjusting original data, screening boundary points with small elevation difference, and calculating the water level and water surface relation of a lake; and performing statistics on water surface areas under different water levels to obtain a water level area change sequence. According to the method, through cooperative adjustment and automatic fusion based on multi-source spatial data, through elevation correction, spatial anomaly screening and time sequence change tracking, the consistency of spatial data is improved, elevation anomaly information is dynamically corrected, the discrimination capability of a water body spatial distribution boundary is enhanced, and dense water surface change information extraction is realized based on spatial overlapping data; the space node and area change trend under the water level grading condition can be automatically concluded, and the dynamic response characteristics of the lake space structure can be systematically reflected.

The invention provides a bank anti-money laundering data model construction method based on privacy computing, and the method comprises the steps: recognizing node role diversity and topology mutation characteristics through an encryption topology framework, extracting transaction paths and a multi-layer nesting relation between nodes, analyzing the role distribution of a transaction main body, and obtaining a dynamic path mutation set; according to the dynamic path mutation set, accounts with short use cycles and accounts with frequently changed transaction subjects are screened out and marked as potential suspicious transaction subjects, and a preliminary suspicious transaction subject list is obtained; identifying transaction records of the preliminary suspicious transaction subject list, screening out high-risk transaction subjects with false registration information or key identity information missing, and obtaining a high-risk transaction subject subset; and counting a plurality of small transactions of the high-risk transaction subject subset, identifying distribution characteristics and multi-layer transfer paths of the small transactions, and extracting fund splitting modes and hidden flow direction characteristics to obtain a high-risk transaction path characteristic graph.

The invention relates to a multi-dimensional data blood relationship tracking method. The method comprises the following steps: S1, data blood relationship meta-information collection; s2, constructing a multi-dimensional blood relationship; s3, blood relationship analysis; s4, an intelligent early warning mechanism; s5, effect verification is realized; a multi-dimensional data blood relationship tracking system comprises a meta-information acquisition module, a blood relationship construction module, a blood relationship analysis module, an intelligent early warning module and an effect verification module. The method has the advantages of high blood relationship tracking accuracy and high analysis capability.

The embodiment of the invention provides a knowledge graph construction method and device, and the method comprises the steps: carrying out the knowledge point analysis of a target document through a knowledge point analysis model, obtaining a plurality of initial document knowledge points in the target document, carrying out the relation recognition of the plurality of initial document knowledge points, and carrying out the relation recognition of the plurality of initial document knowledge points; obtaining an initial association relationship between the initial document knowledge points; performing knowledge point merging on the plurality of initial document knowledge points to obtain a plurality of target document knowledge points, and performing relationship analysis on the plurality of target document knowledge points and the initial association relationship by using a relationship analysis model to obtain a target association relationship among the target document knowledge points; based on the multiple target document knowledge points and the target association relationship among the target document knowledge points, constructing a target knowledge graph; the problems of low efficiency and poor accuracy in the knowledge graph construction process are avoided, and efficient and accurate construction of the knowledge graph is realized.

The embodiment of the invention discloses a data consanguinity analysis method and device, a storage medium and electronic equipment, and the method comprises the steps: firstly, obtaining structured query processing information corresponding to a first target table to which a first target field belongs for the initial first target field, and then obtaining structured query processing information corresponding to a second target table to which the first target field belongs according to the structured query processing information; judging whether a termination condition is met at present or not, if not, performing upstream source disassembly on the structured query processing information to obtain the latest structured query processing information corresponding to the upper layer, and so on, if yes, constructing a data blood relationship tree, then explaining the data blood relationship tree to generate corresponding prompt information, and sending the prompt information to the upper layer of the structured query processing information to the upper layer of the structured query processing information to the upper layer of the structured query processing information. And inputting the prompt information into the trained data blood relationship analysis large model to obtain a data blood relationship analysis result. According to the scheme, the intricate data blood relationship can be linearized and standardized, and the risk interpretation is humanized.

The invention relates to the technical field of food and beverage, in particular to a food and beverage network sales trend prediction model construction system and method based on multi-source data fusion and deep learning. Comprising a data acquisition unit; a data processing unit; the model construction unit is used for constructing a deep learning prediction model, and an improved LSTM-Transform fusion algorithm is adopted to realize nonlinear mapping modeling of the food and beverage sales trend by integrating time sequence feature modeling and a global dependency relationship analysis technology; a model training verification unit; and a prediction output unit. According to the method, multi-source data such as network sales platform data, social media emotion texts, weather information and industry information are integrated, cross-correlation features such as time dimension features, text emotion features and price elasticity-weather influence are extracted in combination with a feature engineering technology, influence factors of food and beverage sales are comprehensively covered, and the sales quality is improved. The problem that a traditional scheme is single in data dimension is solved.

The invention discloses a multi-source field-level blood relationship tracking method and system. The method comprises the following steps: extracting features from a multi-source heterogeneous system, matching templates, triggering a corresponding acquisition strategy, and outputting an original blood relationship log; reading an original blood relationship log to generate an abstract syntax tree, extracting a source / target field set to output a field-level mapping relation, further writing the field-level mapping relation into a graph database to form a blood relationship edge, writing metadata into a relational database, establishing a two-level index, and outputting a blood relationship graph capable of being quickly positioned; and taking the blood relationship map as input, returning a target path with the maximum sum of confidence coefficients by adopting a weighted shortest path algorithm, outputting a link-level traceability result, and displaying the link-level traceability result. According to the method, complete tracking of the field-level blood relationship is realized, the query response time is shortened to a millisecond level, rapid access to multiple types of data sources is supported, and the problems of insufficient blood relationship analysis precision, low query efficiency and weak multi-source adaptability in the prior art are solved.

The invention discloses a tidal reach water level and flow relation analysis method based on tidal level harmonic analysis. The method comprises the steps of collecting actually measured water level data, basin river topography and digital elevation data of a tidal reach hydrologic station; characteristic parameters and water surface elevation change characteristics of all partial tides of the tide level of the tidal river reach are obtained on the basis of tide level harmonic analysis, and then independent tide flow and river runoff flow are obtained; establishing a tidal reach one-dimensional hydrodynamic model, and analyzing a flow change process and a water level flow change relation under different water level change scenes; and analyzing the contribution degree of the tidal flow and the river runoff to the actual flow and the change process by utilizing a multivariate regression analysis and drift diameter analysis method to obtain the change characteristics of the water level and flow relationship of the tidal reach. The method has the advantages that the water level and flow relation of the tidal reach can be rapidly analyzed, complex estuary hydrodynamic process simulation is not needed, and the method is also applicable when actually measured flow data is lacked.

The invention provides a call center dialogue sentiment analysis method and system fusing voice and text, and relates to the technical field of voice signal processing, and the method comprises the steps: obtaining a voice signal of a call center and a corresponding transliteration text, extracting intonation, speed, sound intensity and pause sentiment features from the voice signal, meanwhile, deep language analysis is carried out on the transliterated text, and semantic emotion features related to context are extracted; and performing cross-modal correlation analysis on the voice emotion features and the semantic emotion features, and generating a time sequence correction coefficient for feature alignment by constructing a corresponding relation analysis framework between feature sequences. According to the method, language emotion information in voice and text is integrated, more accurate and comprehensive recognition of conversation emotion of the call center is realized, customer satisfaction is improved, and reliable emotion analysis basis is provided for efficiently processing customer appeals.

The invention relates to the technical field of artificial intelligence, in particular to a government affair service-based machine learning intelligent question and answer analysis method and question and answer system. The method comprises the following steps: performing government affair service data acquisition based on a multi-channel government affair service data acquisition engine to obtain government affair service data; based on the government affair service data, carrying out government affair service question-answer associated combination matrix design, and generating a government affair service question-answer combination matrix; performing hierarchical topic relationship analysis of question and answer combination on the government affair service question and answer combination matrix to generate question and answer combination hierarchical topic relationship data; designing a government affair service intelligent question-answering model through question-answering combination level theme relation data; receiving instant government affair service user question data; and transmitting the instant government affair service user question data to the government affair service intelligent question and answer model to carry out government affair service multi-answer intelligent feedback processing so as to obtain government affair service multi-answer intelligent sorting data. According to the invention, efficient and accurate government affair service question and answer intelligent feedback is realized.

The invention provides a code quality adaptive evaluation and optimization method, system and device and a medium, and belongs to the technical field of computers. The method comprises the following steps: performing redundant information cleaning and standardization processing on an original code; selecting an adaptive parser to convert the processed code into an abstract syntax tree, and optimizing the abstract syntax tree; performing embedding processing on each node of the abstract syntax tree by using a pre-trained deep learning model to generate a node semantic vector, and performing aggregation, feature extraction and vector normalization operation to obtain a semantic vector representing code semantic features; extracting a dependency relationship from the abstract syntax tree to construct a program dependency graph, extracting node features by using a neural network model, and generating a dependency relationship analysis result; based on the semantic vector and the dependency analysis result, identifying a code potential problem by using an optimization algorithm and generating an optimization suggestion; codes are automatically modified according to optimization suggestions, the optimization effect is re-evaluated, and related models and algorithms are iteratively optimized according to feedback information.

The invention discloses a tea tree liquid phase chip and application thereof, and relates to the technical field of molecular detection. The invention discloses a tea tree liquid phase chip and application thereof, according to site screening requirements and probe design principles, the tea tree liquid phase chip comprises 5781 SNP sites, and tea tree resource genetic typing can be realized based on a target interval genome sequence liquid phase capture accurate positioning sequencing typing technology. The tea tree liquid chip can realize low-cost genetic typing, is mainly specific to tea trees, can realize variety identification and genetic relationship analysis of the tea trees, scientifically guides hybridization improvement work of the tea trees and assists protection and development of germplasm resources of the tea trees, and has relatively high application values in multiple fields of tea tree breeding. The tea tree liquid phase chip can be used for tea tree genetic diversity evaluation, germplasm resource and genetic relationship identification, genetic map construction and gene localization, whole genome association analysis and tea tree molecular marker assisted breeding.

The invention discloses a short-term wind power cluster power prediction method supporting conventional meteorological typing and extreme weather small sample expansion, and the method comprises the steps: comprehensively utilizing NWP data, wind power plant operation data and other related information; the technologies such as a Granger causal relationship analysis algorithm, a hierarchical clustering method and a neural network are utilized to realize conventional meteorological scene typing and extreme weather small sample amplification, so that the defects caused by fuzzy prediction scene, low sensitivity to meteorological changes and insufficient consideration to extreme conditions in traditional wind power prediction can be overcome; and moreover, the precision of the wind power short-term prediction model can be improved, and the problems that the prediction model cannot cope with the influence of extreme weather and the power prediction result still has large deviation are solved.

The invention discloses an unmanned aerial vehicle mine resource monitoring system based on multi-modal data fusion, and relates to the field of mine resource monitoring, and the unmanned aerial vehicle mine resource monitoring system comprises a model construction module, an adjustment control module, an atlas construction module, a data fusion module and a monitoring module. A function is introduced for analysis, an unmanned aerial vehicle motion control strategy is obtained; a model is used for entity recognition; a relation triple is obtained through dependency syntactic analysis; coordinate atlas space attributes are fused, a topological relation is constructed, a conflict relation is verified, and implicit relation analysis is completed; the method comprises the following steps: obtaining a mine modular geological map architecture, generating node embedding features by using map attention network coding, associating sensor data with map nodes by using a cross-modal attention mechanism, outputting physical enhancement features, executing monitoring analysis on mine resources, keeping stable unmanned aerial vehicle group control, and ensuring that fused data is accurate.

The invention relates to the technical field of pattern classification, in particular to a pattern classification method and system based on a regional label, and the method comprises the following steps: extracting a local label threshold based on a pixel gray range, analyzing the regional difference of a pattern image, screening a change optimal group as a reference, analyzing a gradient direction and amplitude mutation, and screening layered breakpoints. And integrating the regional features, and judging category attribution to obtain a discriminant quantity. According to the method, detailed pixel relation analysis and regional structure feature extraction are executed step by step, label affiliation accurate adjustment is achieved based on local statistics and spatial attribute association synchronization, boundary anomaly is judged through gradient and label change multiple parameters, layering is carried out, nodes are automatically positioned according to attribute mutation, and category affiliation is decided by multi-dimensional attribute interaction. Layered and partitioned analysis of high-detail and complex patterns is realized, hierarchical attribution requirements under multiple scenes are adapted, the region classification accuracy is enhanced, and the response capability to the internal differential and hierarchical relationship of the pattern structure is improved.

The invention relates to the technical field of communication network security, in particular to a communication network security auditing method and system under multi-source heterogeneous data association, and the method comprises the steps: collecting network flow data and log data in a communication network in real time; distinguishing non-service flow data and signaling data (also called service flow data) from the collected network flow data; performing deep analysis and classification processing on non-service flow data, performing unified analysis and association analysis on log data, and performing protocol decoding and content analysis on signaling data; constructing a multi-source heterogeneous data association relationship analysis model, integrating non-service flow data, signaling data and log data, mining space-time association and causal relationship among the data by applying a machine learning algorithm, and constructing a network security situation map; and based on a multi-dimensional analysis result, automatically generating a security audit strategy matched with the multi-dimensional analysis result. According to the invention, accurate detection, real-time blocking and risk early warning of communication network security threats are realized, and the network security protection capability and the operation and maintenance efficiency are improved.

The invention belongs to the field of big data, and relates to a data consanguinity analysis visualization method, which comprises the following steps of: obtaining metadata of target data from a plurality of data sources, and determining a metadata type and a metadata attribute of the metadata; obtaining a blood relationship model of the target data; based on the blood relationship model, the metadata type and the metadata attribute, performing data blood relationship analysis on the target data by adopting a preset graph algorithm to obtain a blood relationship analysis result of the target data; and based on the blood relationship analysis result, establishing a blood relationship graph of the target data, and displaying the blood relationship graph in a preset interactive interface. The invention further provides a device, equipment and a medium. In addition, the invention also relates to a block chain technology, and the metadata can be stored in a block chain. The method can be applied to the fields of financial insurance, medical health management and the like, and the blood relationship graph of the data can be efficiently and accurately established and displayed.

The invention provides a large model retrieval enhancement generation method fused with a multi-source knowledge base, and relates to the technical field of large model retrieval enhancement generation. The method comprises the steps that firstly, user query is analyzed, multiple implied independent retrieval intentions are recognized and separated, and a knowledge source identifier is distributed to each retrieval intention subtask based on a mapping rule; according to the method, related evidence fragments are acquired in parallel from a heterogeneous knowledge source library in a targeted manner, multi-level relation analysis is further performed on the original evidence fragments by introducing a cross validation engine, so that a structured evidence conflict graph is constructed, and then the evidence conflict graph is intelligently processed by applying a game theory excitation algorithm. The method comprises the following steps: obtaining an evidence set with internal consistency, forming a guide signal through structured analysis, and finally, inputting original user query, the synthesized evidence set and the guide signal into a large language model in combination with a pre-constructed cue word template to generate a final answer under constraint, so as to realize retrieval enhancement generation fused with multi-source knowledge base retrieval.