2371 results about "Confidence score" patented technology

In a described embodiment, a system for information processing is provided including a data acquisition module configured to receive feedback corresponding to one or more outputs generated by a language model. The system further includes a cognitive reasoning module configured to evaluate the reasoning process of the language model, emulate cognitive functions including metacognitive processes, and generate an assessment based on an analysis of the received feedback, wherein the assessment includes classifying the one or more outputs into components, assigning quality scores for each component, and identifying an improvement corresponding to the one or more outputs. Additionally, the system includes a process adjustment module coupled to the cognitive reasoning module for adjusting the reasoning process of the language model based on the assessment is provided. A refinement module coupled to the process adjustment module is provided for iteratively refining the reasoning process based on subsequent updates to the generated assessment until a performance threshold is met.

A computing system for adaptive electronic message classification employs a multi-agent architecture comprising a media feature analysis system, a user context refinement system, and a response synthesis system. The media feature analysis system generates pillar scores including message type, intent, and link risk scores with associated confidence values using trained classification models. When pillar scores and confidence values do not satisfy predetermined threshold conditions, the user context refinement system dynamically constructs contextual prompts using the pillar scores and confidence values as input parameters. User responses generate score modification data that refines the pillar scores and contextual response data for recommendation generation. The response synthesis system generates refined classifications and personalized recommendations using the refined pillar scores and contextual response data. An orchestration system coordinates agent interactions using learned uncertainty points and implements asymmetric influence algorithms with variable weighting based on content and URL analysis concordance.

The invention discloses an electric power field knowledge question-answering system construction method based on a large language model, and relates to the field of electric power field knowledge question-answering, and the method comprises the steps: judging the data type of electric power field knowledge, and carrying out the processing of the electric power field knowledge according to the judgment result through matching with a processing technology, and generating an entity relation triple; constructing a power field knowledge graph; optimizing the power field knowledge graph based on the attention network, outputting an answer causal path of the fault problem by using the optimized power field knowledge graph, and marking a confidence score of the answer causal path; and inputting the solution causal path and the confidence score into a language model to obtain a fault question answering result, and optimizing the question answering result according to the consistency of the fault question answering result and the power field knowledge graph. According to the method, on the premise that the fault diagnosis logic is rigorous and the result is traceable, knowledge in large-scale unstructured literatures in the power industry is activated, so that accurate question and answer services can be provided for operation and maintenance personnel in real time.

The invention discloses a health degree evaluation system and method based on a photovoltaic string, and belongs to the technical field of photovoltaic operation and maintenance intellectualization, and the method comprises the steps: firstly collecting SCADA operation data of a plurality of strings of a photovoltaic power station, including voltage, current, assembly temperature, environment irradiance and power factors; secondly, constructing a whole-station feature reference model, and generating a state vector fusing environment and load characteristics through a nonlinear projection algorithm; aiming at the target group string, extracting a health reference track, calculating a disturbance propagation factor, and identifying a health abnormal state; when the deviation degree exceeds a threshold value, dynamically screening heterogeneous reference group strings, constructing a health score distribution model, and further calculating a confidence score and a prediction residual error; if the score is low and the residual error is unstable, determining that the string is in a sub-health or hidden fault state, and generating an intervention instruction; the method has high accuracy, strong adaptability and good interpretability, and can be widely applied to intelligent diagnosis and refined operation and maintenance management of the photovoltaic power station.

The invention provides a root cause positioning method which can be applied to the technical field of artificial intelligence. The root cause positioning method comprises the following steps: acquiring data in a configuration management database, a network topology tool, a monitoring system and a work order system to form a multi-source heterogeneous data set; performing knowledge extraction on the multi-source heterogeneous data set, extracting equipment attributes, network topological relations, fault event entities and timestamps, and storing the equipment attributes, the network topological relations, the fault event entities and the timestamps as structured knowledge; mapping real-time index data in the structured knowledge into dynamic attributes of an entity, and constructing a dynamic knowledge graph; based on a graph neural network and in combination with time sequence features, learning a time sequence dependency relationship and a propagation path between fault events in the dynamic knowledge graph; and outputting a root cause entity, a confidence score and a fault propagation path of the fault event through a causal inference algorithm in combination with the multi-dimensional evidence. The invention further provides a root cause positioning device and equipment, a storage medium and a program product.

The invention discloses a multi-modal data driven general report generation method and system based on a large model, and belongs to the technical field of intelligent report generation. Firstly, texts, images and sensor data related to a report theme are obtained and subjected to standardized preprocessing; analyzing the report generation instruction, and matching and querying a task modal mapping library matching modal configuration scheme according to a task demand; quantitatively evaluating the data quality of each modal, dynamically calculating the final decision weight of each modal in combination with the basic weight, and distributing the final decision weight to a corresponding processing path to form dominant, supplementary and reference data; inputting the dominant data and the supplementary data into a multi-modal model for analysis to obtain a preliminary conclusion with confidence score, performing consistency judgment, if no conflict exists, performing fusion to form a comprehensive conclusion, and if the conflict exists, combining a quality evaluation result and an arbitration rule to complete conflict judgment; and finally, inputting the comprehensive conclusion and the reference data into a large language model to generate a report text, and outputting a complete report after typesetting and proofreading.

The invention discloses a medical health question and answer method and device based on an intelligent agent and a medium, and relates to the technical field of artificial intelligence medical question and answer, and the method comprises the steps: receiving an intention analysis report through a master control intelligent agent, accessing a dynamic dialogue context pool, and generating an intelligent agent cooperation instruction according to the state of the intention analysis report and the state of the dynamic dialogue context pool; performing credibility evaluation on the preliminary medical answer report by using a credibility calibration agent, generating a confidence score and an evidence conflict level, and performing multi-dimensional weighted fusion and risk mode recognition by using a dynamic risk evaluation strategy to generate a diversified disposal instruction; and when the diversified treatment instruction is issuing permission, performing safety compliance check on the preliminary medical answer report to generate compliance medical answers. According to the invention, multi-dimensional control of credibility and security of medical answers is realized, and finally the beneficial effects of providing personalized medical questions and answers and ensuring that information is real, reliable, compliant and safe are achieved.

The systems and methods disclosed herein orchestrate task execution among autonomous (or semi-autonomous) AI agentic models (“agents”) responsive to a received query by using a hierarchical model cascade to classify queries into agent domains. Queries are processed iteratively by a series of hierarchical levels containing one or more AI models, where each layer is more complex and imposes fewer resource constraints. Each level generates a classification and a confidence score pertaining to the classification. A dynamic bypass mechanism analyzes the classifications and confidence scores at each level to dynamically determine if one or more levels of the hierarchy can be bypassed while resulting in an accurate classification. The final classifications are matched to one or more agents that process the query. Responses from the candidate agents are aggregated into an output that is responsive to the input.

The invention relates to the technical field of pipeline inspection, in particular to a defect identifying and positioning method. Comprising the following steps: generating a uniform node feature tensor through coordinate mapping and feature fusion by synchronously collecting a pipeline inner wall image, an ultrasonic echo and an electromagnetic eddy current signal; constructing a space-time heterogeneous feature graph, integrating three types of relationships of a space adjacent edge, a time evolution edge and a semantic similarity edge, and dynamically optimizing a graph structure by utilizing a trainable fusion factor; a heterogeneous edge decoupling convolution and dynamic attention mechanism is designed, space-time semantic features are extracted through channels, neighborhood information is aggregated, and high-resolution defect classification is achieved; based on a classification result and a residual tensor of an original feature, a defect space position is accurately predicted through a coordinate inversion network, and positioning robustness is improved by combining positioning confidence score and weighted aggregation; and finally fusing the equipment track and the pipeline three-dimensional model to realize defect geographic coordinate mapping and interactive visualization. According to the method, the defect identification precision and the positioning reliability in a complex pipeline environment are remarkably improved.

The invention discloses a multi-source data fusion type intelligent data management system, which comprises a data preprocessing module used for accessing a heterogeneous data source and generating a standardized data frame; the cognitive sub-graph construction module is used for generating cognitive sub-graphs and forming a cross-source evolutionary multi-cognitive hypergraph; the event aggregation module is used for aggregating entities and relationships according to event fingerprints and outputting a multi-source observation chain; the causal checking module is used for performing causal comparison and conflict detection and outputting a checking result, a positioning report and a treatment suggestion; the fusion updating module is used for distributing weights and performing weighted fusion, outputting fusion representation and confidence score, recording conflicts and updating the hypergraph; and the memory pool module is used for storing the hypergraph, the check result and the fusion representation, executing weight adjustment, forgetting and elimination, and outputting a service interface. According to the method, intelligent fusion and dynamic management of multi-source heterogeneous data are realized by constructing a cross-source self-evolution multi-cognitive hypergraph and an active evolution type fusion memory pool.

Systems, methods, and devices that relate to intelligent query decomposition and parallel routing for specialized model processing are disclosed. In one example aspect, the system receives a query from a user comprising a request relating to a particular domain. The system determines, using a decomposition model, a set of sub-queries based on semantic boundaries, syntactics, tasks, relationships, and rules relating to particular domains. The system inputs the set of sub-queries into a routing model to determine a set of specialized models. For each sub-query, the system routes the sub-query to a respective specialized model, generates an output, and assigns a confidence score. The system detects conflicts among outputs using a conflict detection model configured to identify discrepancies. The system generates an aggregated output by combining outputs according to a weighted aggregation algorithm prioritizing higher confidence scores and conflict resolution rules, then displays the aggregated output.

The invention discloses an information retrieval system and method based on semantic normalization, and relates to the technical field of artificial intelligence information, and the method comprises the steps: collecting a semantic query record input by a user, carrying out the preliminary semantic analysis, and generating structured data; on the basis of the structured data, entity disambiguation is carried out by utilizing a knowledge graph, abstract classes are generated through a neural network, calibration and dynamic weight adjustment are carried out, and high-confidence entity abstract classes and confidence scores are generated; entity abstract classes and confidence scores are combined with user contexts, an action-value function is calculated through a value network, and an optimal action is selected by utilizing a-greedy algorithm; executing semantic normalization mapping according to the optimal action, and obtaining an intermediate expression by using a meta-symbol dynamic generator; and performing index retrieval and multi-dimensional sorting based on the intermediate expression to generate a sorted retrieval result list. According to the method, the semantic fragmentation problem of multi-modal query is solved, and deep semantic alignment and dynamic weight calibration of heterogeneous data are realized.

The invention relates to an RAG-based voucher classification method, a medium and equipment, and the method comprises the steps: receiving digital image data of a to-be-classified voucher, carrying out the multi-modal optical character recognition processing to generate a structured OCR result, extracting a text semantic feature vector and a visual layout feature vector based on the structured OCR result, carrying out the fusion of the text semantic feature vector and the visual layout feature vector to generate a multi-modal query vector, and carrying out the classification of the to-be-classified voucher. Similar samples and semantic similarity scores and category metadata thereof are obtained through approximate nearest neighbor retrieval, after an initial candidate category list is generated, key field values are extracted for each candidate category, evidence credibility scores are calculated, comprehensive confidence scores are generated by fusing the semantic similarity scores and the evidence credibility scores, reordering is conducted, and a candidate category list is obtained. And finally, selecting a classification decision path according to the score distribution, and outputting a classification result and an interpretability report. The accuracy and robustness of voucher classification are effectively improved, and complex voucher scenes with changeable formats and fuzzy semantics can be processed; and the interpretability and reliability of the classification decision are enhanced.

The invention discloses a knowledge graph generation method and system based on a large model, and the method comprises the steps: carrying out the preprocessing of data, obtaining a data set, and determining a target mode of a knowledge graph; extracting a triple conforming to the type of the target mode from the data set to obtain a candidate knowledge triple; carrying out multi-dimensional verification on the candidate knowledge triad to generate a confidence score; performing consistency verification and conflict resolution on the candidate triad set with the confidence score higher than a preset threshold value to obtain a verification candidate knowledge triad; performing entity linking and relationship standardization processing on the verification candidate knowledge triples, and mapping the verification candidate knowledge triples to a knowledge graph of a corresponding target mode to obtain standard candidate knowledge triples; and fusing the standard candidate knowledge triples into the knowledge graph database, and updating the knowledge graph database. According to the method, the powerful natural language understanding and knowledge reasoning potential of the LLM can be utilized to the maximum extent, and the inherent defects of the LLM are actively and systematically overcome by introducing an innovative mechanism.

The invention provides a patient vital sign anomaly detection method based on an artificial intelligence technology, and relates to the technical field of data processing, and the method comprises the steps: collecting original sign data of a patient; calculating multi-dimensional characteristic parameters; establishing an individual baseline model, and determining a comprehensive reference interval in the model; the vital sign features monitored in real time are constructed into multi-dimensional feature vectors, the multi-dimensional feature vectors are input into the individual baseline model, and the deviation degree of the real-time multi-dimensional feature vectors in the comprehensive reference interval is calculated; performing a clustering analysis to identify an anomalous aggregation region; performing trend analysis, calculating change direction consistency and continuous change amplitude of the multi-dimensional feature vector, and generating a trend analysis result; calculating an accumulated change index in a continuous time window according to a trend analysis result to obtain a dynamic confidence score; when the dynamic confidence score continuously exceeds an adaptive threshold value, determining that a vital sign abnormal event exists; the autonomy and accuracy of the method for detecting the vital sign abnormity of the patient are improved.

The invention discloses a large-model knowledge retrieval method based on knowledge graph enhancement, which comprises the following steps: S1, collecting multi-source data, constructing an original knowledge triple set, and generating a knowledge graph; s2, inputting natural language query data, executing semantic analysis and intention recognition, and forming a query semantic vector and a structured query template; s3, performing combined convolution operation on the target entity node and the relation vector through an improved combined graph convolution network CompGCN to generate a candidate knowledge list and distribute confidence; s4, calculating the similarity between the query vector and the knowledge vector by utilizing a semantic embedding mechanism and DistilBERT acceleration to obtain a first retrieval result set and a credibility score; s5, fusing the two types of retrieval results to obtain a knowledge result set; and S6, generating an answer text according to the fused knowledge result set, and outputting a reasoning path, entity reference and a relation link. According to the method, the accuracy, the response speed and the knowledge reasoning transparency of complex query processing are improved.

The invention discloses an electricity larceny identification method based on graph calculation, and particularly relates to the technical field of electricity utilization anomaly detection of an electric power system. Historical power consumption data and a power supply topological relation of power consumers are collected, and a multi-dimensional behavior graph model fusing behavior characteristics and structural information is constructed; performing structure disturbance analysis on each node in the graph, calculating information entropy change before and after node removal, performing attention fusion on a time sequence behavior feature of the node and a structure disturbance vector, constructing a joint feature vector, and enhancing feature expression through spectral clustering and linear reconstruction; a behavior propagation field and a disturbance adjustment mechanism are introduced into the graph to form a disturbance response graph, and an abnormal gathering area is identified through path energy analysis and focusing area fitting; calculating confidence scores of the nodes and outputting a suspicious user list; the method can realize efficient identification of electricity stealing behaviors with strong concealment and complex transmissibility, and has the advantages of high precision, strong interpretability and wide application scene adaptability.

The invention discloses an electric power system safety early warning method and system based on multi-modal cooperation, and relates to the technical field of electric power system safety early warning, and the method comprises the steps: collecting multi-source operation data, carrying out the preprocessing, carrying out the multi-modal feature extraction and fusion based on the preprocessed data, and carrying out the multi-modal feature extraction and fusion. Inputting an edge detection model and outputting an abnormal confidence score in combination with an attention mechanism; and performing alarm grading according to the abnormal confidence score, constructing a causal diagram for alarms with high risk levels in combination with associated security events, and performing future attack path prediction by adopting a time sequence diagram neural network. According to the method, multi-scale convolution and a channel attention mechanism are fused, the extraction capability of the multi-source data time sequence features of the power system is enhanced, the anomaly detection precision is improved, dynamic attack path prediction is realized in combination with RMTPP and causal atlas topological constraints, sequence modeling is enhanced through self-attention and position coding, and the detection accuracy is improved. And the perspectiveness and the reliability of the safety early warning of the power system are obviously enhanced.

The invention relates to the technical field of material management, and discloses a dynamic rule generation and self-adaptive auditing system and method for material management, and the system comprises a rule intelligent extraction module, a rule management knowledge base, an enhanced auditing engine, a man-machine cooperation calibration module and a self-adaptive execution module. The method comprises the steps of automatic rule extraction, rule storage and management, enhanced auditing and reasoning, man-machine collaborative calibration, knowledge base real-time optimization and adaptive routing execution. According to the method, the rule is automatically extracted from the unstructured document, the problem that a traditional system rule depends on manpower and is lagged in updating is solved, dynamic optimization of the rule and confidence is achieved by introducing a man-machine collaborative feedback closed loop, the accuracy and transparency of an audit decision are improved by enhancing reasoning and explainable decision technologies, and the audit efficiency is improved. And the optimal balance between auditing efficiency and risk control is realized through self-adaptive routing execution based on credibility.

Artificially intelligent systems and methods for financial coaching provide personalized, fiduciary-compliant financial guidance through advanced machine learning architectures with measurable performance criteria. The systems implement privacy-preserving processing pipelines that detect personally identifiable information using multi-layered pattern recognition including regular expressions for formatted data sequences, named entity recognition with confidence thresholds above 0.85, and contextual analysis algorithms. A multi-step artificial intelligence processing workflow includes automated language detection, emotional tone classification with confidence scoring, financial profile transformation using predefined templates, context-aware question rephrasing, and semantic similarity matching employing vector embeddings with financial domain vocabulary weighting applying multiplier values between 1.3-2.0. Specialized training methodologies expand datasets through mathematical transformation functions utilizing statistical standard deviations with incremental variations between 0.5-2.0. Mood-based escalation logic automatically transfers users to human advisors when emotional indicators exceed confidence thresholds above 0.8. The systems maintain response times below 5 seconds while providing regulatory compliance through curated content sources and predefined fiduciary instruction parameters.

The invention relates to the technical field of communication, and discloses a communication state monitoring management method and system. The method comprises the following steps: acquiring a transmission data packet in a communication network in real time, extracting protocol header information and load content of the data packet, and generating an original monitoring data sequence containing a protocol type, a timestamp and load characteristics; and performing multi-dimensional correlation analysis on the original monitoring data sequence, identifying a logical relationship between protocol header information and load content, detecting an abnormal correlation mode, and generating an abnormal event mark set with a confidence score. And constructing an adaptive behavior model based on the original monitoring data sequence and the abnormal event mark set, and capturing normal behavior characteristics of the communication network by dynamically updating model parameters. Transmitting data packets collected in real time are input into the self-adaptive behavior model for deviation degree calculation, abnormal data packets exceeding a deviation threshold value are screened out, and historical abnormal event mark sets are associated to generate abnormal context records.

A multi-agent collaboration tool for collaborating among multiple large language model (LLM) agents is provided. A problem statement is received from a user. A first LLM agent is automatically selected to provide a first answer to the problem statement as a first confidence score for the first LLM agent is more than a second confidence score for a second LLM agent to provide the first answer to the problem statement. The second LLM agent is automatically selected to provide a second answer based on the first answer to the problem statement as a third confidence score for the second LLM agent is more than a fourth confidence score for the first LLM agent to provide the second answer based on the first answer to the problem statement. A solution to the problem statement is provided based on the second answer.

The invention relates to the technical field of medical prediction, in particular to a Web-based big data driven stroke risk prediction method, which comprises the following steps: S1, multi-scale data acquisition and alignment: acquiring physiological data, behavior data and health state data of a user; s2, causal path identification: generating a causal feature sub-graph with confidence rating; s3, feature fusion modeling: performing response modeling on fluctuation features in the real-time physiological data, performing trend feature extraction on behavior data, and dynamically allocating fusion weights of the fluctuation features and the trend features to generate unified risk characterization features; s4, risk evolution simulation: simulating the influence of different intervention strategies on the stroke risk, and generating a risk evolution trajectory; and S5, intervention scheme generation: generating a personalized intervention scheme. According to the invention, the accuracy of risk prediction is improved, and the change of the health condition of the patient can be responded in real time.

The invention discloses a partition detection method for board card quality, and relates to the technical field of detection. The method comprises the following steps: dividing a board card into functional blocks according to a functional structure, and collecting thermal field distribution characteristic data of each block; constructing a standard thermal template, calculating the thermal imaging matching degree of the functional blocks by using an Euclidean distance algorithm, and marking the thermal imaging matching degree exceeding a preset threshold value as an abnormal thermal response region; collecting resistance and resistance gradient change data of an abnormal thermal response area, judging the area as a defect cluster area if conditions are met, and calculating a resistance abnormal amplitude; integrating defect cluster region data to form a composite characteristic spectrum, inputting a defect sample matching model, and outputting a defect classification result and a confidence score; calculating a regional quality score by combining fuzzy logic reasoning and a neural network algorithm and integrating a thermal imaging matching degree, a resistance abnormal amplitude and a confidence score; and fusing the defect classification result and the regional quality score to generate a quality grade label. And based on the quality grade labels, board card quality partition detection is completed.

The invention discloses an aquaculture comprehensive guarantee method and system based on multi-modal data acquisition, and relates to the technical field of intelligent aquaculture, and the method comprises the steps: collecting aquaculture multi-modal data and an aquaculture image of an aquaculture region, carrying out the aquaculture feature capture of the aquaculture image, and outputting a visual feature group; inputting the visual feature group into a MobileNet lightweight model, and outputting a health state label and a confidence score; based on the health state label and the confidence score, outputting disease type data through a residual network structure optimized by transfer learning; constructing a graph neural network according to the disease type data, and generating a dynamic regulation and control scheme; and the control center executes the dynamic regulation and control scheme, carries out regulation and control effect detection and target comparison on the aquaculture area of the dynamic regulation and control scheme, judges whether the regulation and control effect is effective or not, and generates a feedback regulation instruction. According to the invention, through multi-modal data fusion and intelligent decision closed loop, dynamic and accurate guarantee of aquatic product health management is realized.

The invention discloses a traffic image labeling method and device based on a D-S evidence theory and a medium, and relates to the technical field of image labeling. According to the method, two different types of real-time models are simultaneously utilized to process the same image in parallel, and two uncertainty detection results of the same image are obtained; a KM algorithm is utilized to match targets in two uncertainty detection results, then a prior credibility score is utilized to reduce a conflict coefficient, and two different types of small model prediction results are combined through an uncertainty theory. According to the traffic image labeling method and device based on the D-S evidence theory and the medium, the strong generalization ability of the open set model is utilized to make up for the defects of a closed set model, the use of a large model to improve the hardware dependence is avoided, the precision problem of a single small model is made up under the condition that the efficiency is ensured, and the iteration duration of the model is shortened.

The invention discloses a multi-modal power grid fault diagnosis method and system based on a causal event atlas, and belongs to the technical field of intelligent operation and maintenance of power systems. The method comprises the following steps: preprocessing historical fault case data of power grid equipment, and constructing a causal event atlas database; when a fault diagnosis request is received, analyzing the fault diagnosis request by the planning agent, generating an initial fault hypothesis set in combination with power field knowledge, endowing a corresponding credibility score to the fault hypothesis, retrieving the cause subgraph as evidence in an iterative loop, and updating the credibility score of the fault hypothesis by the reasoning agent; and generating a multi-modal diagnosis report after the termination condition is met. According to the method, the problems of insufficient causal modeling and poor interpretability of a traditional method are solved, and the diagnosis accuracy, efficiency and user credibility are remarkably improved.

A condition generator analyzes a set of utility asset images of a particular utility asset using the ML (machine learning) model identify a type and condition of the particular utility asset depicted in the set of utility asset images. The identified condition is assigned a confidence score, and the set of utility asset images includes at least two images of the particular utility asset captured at different angles. The condition generator generates a descriptive tag for the set of utility asset images based on the identified type and condition. The descriptive tag characterizes an operational status of the particular utility asset. The condition generator stores the set of utility asset images and the generated descriptive tag in a utility asset database. The utility asset database stores images of utility assets.

The invention discloses a traffic monitoring video rapid target extraction method for edge equipment, and relates to the technical field of intelligent traffic video processing and edge calculation target detection, and the method comprises the steps: carrying out the adaptive downsampling processing of original video frame data, and generating downsampling video frame data; extracting a foreground target candidate region, and constructing a traffic region-of-interest mask in combination with a lane line detection result; carrying out pixel AND operation on the traffic region-of-interest mask and the foreground target candidate region to generate accurate candidate target region data, and extracting a target feature vector; carrying out weighted fusion on the target feature vector through a lightweight attention mechanism, generating a fusion feature descriptor, and calculating a target confidence score; and carrying out screening and duplicate removal processing on the accurate candidate target area data, and outputting traffic target extraction result data. According to the invention, the target in the traffic video can be rapidly and accurately extracted and processed in a low-delay manner on the edge equipment with limited computing resources.

The invention discloses a knowledge graph completion method based on combined prediction of large and small models, which comprises the following steps: 1, constructing and preprocessing a knowledge graph completion reference data set, and training by adopting a RotatE model to obtain candidate entities generated by the model and confidence scores; 2, constructing a related triad, an adjacent triad and entity long text description based on the query to form a context prompt, inputting the context prompt and the query into a large language model, and performing semantic reordering and scoring by the large language model; and 3, constructing a fine tuning data set, and performing fine tuning on the large language model to obtain the KGC task optimization-oriented large language model. And based on the KGC score, the LLM score and the dynamic weight, outputting a complementation result through joint prediction of a fusion result. According to the method, the structured reasoning ability of the small model and the deep semantic understanding of the large language are effectively combined, the prediction accuracy, robustness and specialty are remarkably improved, and the defects that a single model is weak in generalization ability and insufficient in semantic utilization are overcome.