1279 results about "Similarity matrix" patented technology

The invention discloses an APT attack path reconstruction method based on time sequence diagram comparison clustering and a medium. A security event standardized data set is obtained; mapping each security event into a multi-modal node through a heterogeneous time sequence diagram set construction method, and generating a directed edge to construct and complete a heterogeneous time sequence diagram set; a stage embedding time sequence diagram set is obtained through the joint attack stage set; outputting a time sequence diagram similarity matrix by adopting a multi-scale diagram similarity algorithm of time alignment perception; generating an event semantic sparse matrix based on the threat intelligence knowledge graph; obtaining an image clustering result set; uncertain samples in the graph clustering result set are obtained and processed, and an APT attack path reconstruction result is obtained. The problem that an APT attack path reconstruction method mainly depends on rule matching, single-dimensional feature comparison and manual analysis of security logs or alarm streams is solved, the interpretability of APT traceability is greatly enhanced, and subjective errors of manual research and judgment are reduced.

The invention relates to a hydrological flow prediction method and system based on multi-site time-space association. The prediction method comprises the following steps: carrying out dimension reduction and feature reconstruction on original multi-site hydrological data through an auto-encoder; space-time correlation modeling and adjacency matrix dynamic construction are carried out, a multi-dimensional Euclidean distance matrix between stations is calculated based on a multivariable dynamic time warping (MDTW) algorithm, a similarity matrix is generated in combination with dynamic programming, and a dynamic adjacency matrix is constructed by fusing a geographic space adjacency relation; extracting spatial features of a GCN (Graphics Convolutional Network); carrying out adaptive time sequence decomposition and trend-period modeling; and carrying out multi-stage fusion prediction and result output, and generating a final prediction result through a decoder in combination with the decomposed trend item and periodic item. According to the method, accurate extraction and dynamic correlation modeling of spatial-temporal characteristics of multi-site hydrological data are realized, the accuracy and robustness of single-site flow prediction are improved, and the problems that multi-site spatial-temporal correlation modeling is insufficient, non-linear time sequence alignment is difficult, and single-site prediction precision is limited are solved.

The invention provides a large model training data synthesis method, system and device and a storage medium, and the method comprises the steps: carrying out the preprocessing of original military corpus data, generating a semantic vector through a pre-training language model, dividing the semantic vector into a plurality of topic clusters through clustering analysis, extracting the keyword and representative sample of each topic cluster, forming a theme ontology library; selecting associated topic pairs based on the topic similarity matrix, and generating a synthetic sample through a predefined template; controlling the language style of the synthetic sample through the cue word instruction, and generating a stylized sample; dividing difficulty grades of the stylized samples to obtain graded samples; performing multi-dimensional quality evaluation and screening on the graded samples to obtain screened samples; the screened samples are used for training a large language model, and according to the performance of the large language model in each theme task, theme weights are adjusted, a synthesis strategy is updated, and a sample structure is optimized. According to the method, high-quality and diversified training data can be provided for large model training.

The invention relates to an experimental data processing method and device, an AI analysis module and computer equipment, and belongs to the field of data processing.The method comprises the steps that multi-dimensional original data are partitioned according to types, and formats are unified; generating a similarity matrix based on time and space neighborhood information, and marking abnormal fluctuation points; effective signals are separated through a time-frequency feature matching noise library; extracting multi-layer features of basic statistics, time sequence correlation and trend change; and dynamically screening core features to update the tracking type experimental model. The matched AI analysis module integrates hardware circuits of data partitioning, similarity calculation, anomaly marking, noise matching, feature extraction and model updating, and whole-process acceleration is achieved. According to the method, through multi-dimensional data compatibility processing, accurate anomaly detection, multilayer feature fusion and model adaptive optimization, the experimental data processing efficiency and conclusion reliability are remarkably improved, and the method is suitable for real-time analysis of multiple scenes such as scientific research and industry.

The invention relates to the technical field of calibration tests, in particular to an intelligent metering calibration method and system, and the method comprises the following steps: calculating the similarity between data points, constructing a similarity matrix, further calculating the fitting degree, carrying out the spectral decomposition, extracting a feature vector, carrying out the dimension reduction, mapping to a low-dimensional space, carrying out the data point clustering based on a feature vector matrix, and carrying out the calibration of the data points. An error cluster is identified and an error pattern is determined. According to the method, by calculating the similarity between the data points and mapping the similarity to the low-dimensional space, the error modes can be efficiently recognized and classified, the error correction path can be dynamically adjusted, error correction is more flexible and real-time, the error analysis depth is improved through spectral decomposition and feature vector extraction, and the error analysis accuracy is improved. Therefore, the error trends of the equipment under different conditions are accurately identified, the measurement precision and stability are remarkably improved, the error correction path parameters are updated in real time, the adaptability of the system to the operation change of the equipment is enhanced, and the calibration time is effectively shortened.

The invention provides a conference activity execution task decomposition, arrangement and management method based on an AI technology, and relates to the technical field of conference management, and the method comprises the steps: obtaining conference demand information, converting the conference demand information into digital representation data, employing a bidirectional recursive decomposition strategy to carry out task decomposition, and generating subtask data; calculating a task similarity matrix and a conflict matrix to construct a task execution directed graph to generate initial arrangement data; collecting an execution state to generate feedback data, calculating a risk assessment score, and adjusting a task execution scheme when the risk assessment score exceeds a threshold value. According to the invention, the execution efficiency of conference activities can be improved, the resource conflict risk is reduced, and dynamic optimization adjustment is realized.

A method discloses receiving a plurality of input images, receiving text prompts, generating a visual matrix utilizing the images and an image encoder, generating a text matrix utilizing a text encoder, multiplying the text matrix and the visual matrix to generate an image-text similarity matrix that assigns a numerical value indicating similarities between each of encoded visual descriptors and each of the encoded images, wherein similarities are indicated by entries of the image-text similarity matrix having numerical values that determine a loss function associated with the image-text similarity matrix, identify a gradient of the loss function with respect to parameters associated with the image encoder and parameters associated with the text encoder, utilizing the gradient, update parameters associated with the image encoder or the text encoder, and outputting final updated parameters associated with either the text encoder or image encoder of the machine learning network.

The invention discloses a multi-modal multi-scale cross attention-based target counting method and system, and the method comprises the steps: constructing a similarity matrix based on a block-level visual embedding vector and a text embedding vector; and further establishing comparison learning loss based on multi-classification cross entropy so as to enhance the text pairing capability of the block-level visual vector under the vision-language model. Meanwhile, text embedding containing type semantic information and quantity semantic information and a visual embedding tensor with the same dimension as the text embedding are constructed, an attention score is output through calculation of a zoom dot product model, and finally feature representation of block-level visual embedding in a text embedding vector space is obtained. A multi-scale visual feature interaction module based on a double-branch Cross-ViT structure is constructed, and access and fusion among visual features of different scales are achieved. According to the method, the feature extraction performance of the model on target details and context information of different scales in a complex scene is enhanced, and good visual coding is provided for a downstream density map regression task of target counting.

The invention discloses a time-space state enhanced XGBoost-LSTM-based vehicle position prediction and abnormity early warning method, system and device, and a medium. The method comprises the steps of obtaining a road section traffic flow sequence, constructing an EM clustering model, classifying traffic flow sequence samples in the EM clustering model, and constructing a time state matrix; determining a physical static adjacency matrix, a dynamic steering ratio adjacency matrix and a road section flow dynamic similarity matrix between the road sections, and constructing a spatial state matrix after weighted fusion; an XGBoost-LSTM model is constructed, and model training is carried out; and inputting the constructed time state matrix and space state matrix into the trained XGBoost-LSTM model, and carrying out real-time vehicle position prediction and abnormity early warning. According to the method, the vehicle position prediction precision can be improved, the data acquisition threshold is reduced, the use scene of the prediction method is expanded, and meanwhile, a new dynamic evaluation standard is provided for abnormal event early warning.

The invention provides an active transformer area voltage out-of-limit early warning and photovoltaic bearing capacity evaluation method, which comprises the following steps: acquiring multi-source data of a power distribution transformer area, and generating a standardized multi-source time sequence data set through time alignment processing, data quality evaluation and feature fusion; extracting voltage out-of-limit features based on the time sequence similarity matrix and the node influence matrix; the periodicity, volatility and randomness characteristics of photovoltaic output are analyzed, a coupling relation matrix and a sensitivity vector are established, and an influence evaluation model is constructed; performing out-of-limit early warning based on an early warning index system and a deep learning model; and establishing a multi-dimensional evaluation index system for carrying out bearing capacity evaluation in combination with power flow analysis and the dynamic influence factors. Through multi-source data fusion and multi-dimensional feature analysis, the accuracy of voltage out-of-limit early warning is improved, accurate evaluation of the photovoltaic bearing capacity is realized, and effective technical support is provided for safe operation of the power distribution network.

The invention discloses an intelligent processing method and system for judicial batch filing, and the method specifically comprises the steps: carrying out the cross verification of an original case element through a cross-modal information complementation mechanism, and generating a first case element; constructing a case similarity matrix according to the first case elements, calculating a fact relevance weight between cases by adopting an improved BERT-Siamese network based on the case similarity matrix, synchronously identifying jurisdiction competition and proof conflicts through a legal program conflict detection algorithm, and generating a dynamic grouping scheme; matching a classification document template library according to the dynamic grouping scheme, and calling a Transform-XL enhanced generation model to generate legal documents in batches; and performing full-process intelligent verification on batches of legal instruments through the association map, and detecting material integrity, logic consistency and legal compliance. According to the method, intelligence, high efficiency and precision of judicial batch filing are realized, and the case processing efficiency and the document generation quality are improved.

The invention provides an underground pipe network defect intelligent identification and early warning method and system based on deep learning, and relates to the technical field of pipe network detection, and the method comprises the steps: employing a pipeline detection robot to obtain multi-source data, and carrying out the preprocessing; mapping the features to a manifold space, constructing a similarity matrix, and carrying out geometric difference weighted fusion; constructing an incidence matrix to calculate a spatial distance and mechanical strength, and carrying out iterative propagation to generate a defect characteristic spectrum; and acquiring an evolution sequence by adopting a self-adaptive sliding window, determining influence factors based on anti-fact intervention, and calculating a state transition probability to determine an optimal maintenance scheme. According to the method, the pipe network defect identification accuracy is improved, and pipeline state prediction and maintenance decision intelligence are realized.

The invention relates to an article sequence recommendation method and system based on collaborative filtering and an LLM perspective, and the method comprises the steps: obtaining a user historical data set and an article data set, and carrying out the preprocessing, and obtaining an article title similarity matrix, a user behavior sequence, and historical collaborative filtering interaction information; historical collaborative filtering interaction information is enhanced through a large language model; constructing a sequence recommendation model, inputting the user behavior sequence into the sequence recommendation model for training, and correcting the deviation of the user behavior sequence through comparative learning according to the item title similarity matrix and the enhanced historical collaborative filtering interaction information to obtain a trained sequence recommendation model; and inputting the user behavior sequence of the user into the trained sequence recommendation model for prediction, calculating prediction scores, and generating a recommendation list according to score sorting, thereby completing article sequence recommendation. According to the method, the performance of the sequence recommendation system is greatly improved by solving the cold start problem.

The invention discloses a multi-stage spatial-temporal clustering method and system based on a fused mahalanobis distance. The method comprises the following steps: acquiring a spatio-temporal data set, and determining a spatio-temporal neighbor relation of samples in the data set; calculating a space communication distance and a time decay distance of the sample; fusing the space communication distance and the time decay distance by using a mahalanobis distance to obtain a relative distance of the sample; selecting a class cluster center from the data set according to the local density of the sample and the relative distance; adopting a multi-stage distribution strategy to distribute non-class-cluster center samples to corresponding class clusters; wherein the multi-stage allocation strategy comprises an inevitable allocation stage based on space-time shared neighbor and a similarity allocation stage based on a weighted similarity matrix. According to the method, the key problems that an existing space-time clustering algorithm is insufficient in space-time attribute coupling processing and sensitive to distribution errors are solved, and the clustering accuracy, robustness and practicability in the fields of intelligent traffic analysis, seismic sequence recognition and the like are remarkably improved.

The invention provides an open set cross-domain hyperspectral classification method and system based on multi-mode and optimal transmission, and relates to the technical field of hyperspectral image classification. The method specifically comprises the following steps: acquiring and utilizing an original hyperspectral image to construct a labeled source domain data set and an unlabeled target domain data set; utilizing a spatial-spectral feature extractor to respectively obtain spatial-spectral features of the source domain data and spatial-spectral features of the target domain data; and determining a source domain prototype, calculating a similarity matrix between the target domain data and the source domain prototype, solving an optimal transmission matching matrix, distributing a pseudo tag for each target domain data, and performing text correction on all the source domain data and all the target domain data by using a contrast learning loss function to obtain the target domain data. And the source domain prototype and the spatial spectrum feature extractor are optimized by constructing the total loss of minimizing hyperspectral image classification, so that classification prediction can be carried out on the hyperspectral image to be classified, and target domain data can be classified more accurately.

The invention discloses an intelligent storage system supporting multi-source heterogeneous data fusion management, which relates to the technical field of multi-source heterogeneous data fusion, and comprises a data source adaptation module, an intelligent data source adapter is configured in a DataWorks data integration module, multi-source heterogeneous data is accessed through the intelligent data source adapter, and the data source adaptation module is connected with the DataWorks data integration module; extracting a semantic feature vector and a technical feature set; the neural symbol hybrid inference module inputs the semantic feature vector and the technical feature set into a neural symbol hybrid inference engine, calculates a similarity matrix between multi-source heterogeneous data fields through a BERT-based neural network, and imports the technical feature set into a field knowledge graph constructed by Neo4j for symbol logic verification to generate a unified metadata model; according to the method, a semantic similarity matrix calculated by a BERT-based neural network is combined with symbol logic verification of a Neo4j knowledge graph through a neural symbol hybrid inference engine, so that automatic semantic alignment and logic consistency verification of multi-source heterogeneous data are realized.

The invention relates to the technical field of automation and artificial intelligence crossing, in particular to an industrial monitoring and early warning method and system based on a multi-modal large model, and the method comprises the steps: obtaining multi-modal industrial data, carrying out the initial feature processing of the obtained multi-modal industrial data, and carrying out the initial feature processing of the obtained multi-modal industrial data; and performing deep feature extraction fusion based on the extracted preliminary features, taking a multi-modal fusion feature matrix as input to construct a multi-modal large model, performing model training on the constructed multi-modal large model, and performing real-time monitoring and abnormity early warning by using the trained multi-modal large model. According to the method, space-time alignment and feature association are realized through a defect feature matrix of a visual mode, a trajectory similarity matrix of a motion mode and a time sequence reconstruction matrix of a sensing mode and through dynamic time warping and cross-modal parameter sharing; and associated abnormalities of product quality, equipment actions and running states can be captured at the same time, so that the problem that the collaborative analysis capability of a traditional method on multi-source heterogeneous data is insufficient is solved.

The invention provides a generative adversarial-driven intelligent security defense method and system, and solves the problem of dynamic network security defense through three-layer architecture innovation: 1, data fusion layer reconstruction: employing a multi-modal feature extraction engine driven by an MoE architecture, dynamically allocating computing power resources to a plurality of expert models, and improving the heterogeneous data distillation efficiency; an LLM for fine adjustment in the security field is introduced, a cross-modal semantic similarity matrix is constructed, and the accuracy of unstructured threat intelligence analysis is improved; a second dynamic attack and defense layer is constructed, a GPT-4 architecture attack generator is deployed, and generation of a multi-stage APT attack chain is simulated; a double-agent reinforcement learning framework is designed, and the confrontation training efficiency is improved; upgrading a three-cognitive decision-making layer, constructing a dynamic threat map based on a time sequence diagram neural network, and updating an adjacent matrix in real time; a plurality of agent clusters are deployed, the capabilities of encrypted traffic analysis and attack blocking are improved, and the problems of data layer defects, attack and defense confrontation limitation and decision-making layer bottleneck in the prior art are solved.

The invention discloses a supply chain production scheduling optimization method and system based on multi-source data. According to the method, supplier purchase data is extracted through a data crawler, raw material sharing is analyzed in combination with a knowledge graph, and a raw material sharing list is generated; on the basis of delivery records and product attributes, a quality stable value is predicted through linear regression, process similarity is calculated through a text similarity algorithm, and a process similarity matrix is constructed; in combination with historical sales data, a market demand prediction model is constructed by applying multiple linear regression and time sequence analysis. A complex relation among raw material sharing, process similarity and market demands is described through a graph neural network, a product association network graph is generated, an optimal balance point is found between delivery time and a resource utilization rate through Pareto optimization, and a production scheduling scheme is formed; and finally, integrating the performance score, the quality stable value, the raw material sharing list and the production scheduling scheme through an integer programming algorithm, and optimizing a supplier combination. According to the invention, the production scheduling efficiency and precision are improved, and complex business requirements are met.

The invention discloses a tensor depth semi-supervised learning method for high-dimensional small sample data classification. The tensor depth semi-supervised learning method comprises the steps of preprocessing original high-dimensional small sample data; constructing a deep neural network comprising a feature extraction module and a classifier module; constructing a second-order similarity matrix and a third-order similarity tensor based on the low-dimensional embedding representation obtained by pre-training; combining the second-order similarity matrix and the third-order similarity tensor to construct an objective function containing multi-order smooth constraints; then, taking the low-dimensional embedded representation obtained by pre-training as input, performing iterative optimization on the target function by adopting a gradient descent algorithm through a label propagation network formed by a full connection layer, and generating a pseudo label; and inputting the original high-dimensional small sample data, the low-dimensional embedded features and the pseudo labels into the deep neural network, iteratively updating the network in a semi-supervised mode until convergence, and outputting a final prediction result. According to the method, more accurate label propagation is realized, and the semi-supervised classification precision is improved.

The invention discloses a multi-modal offshore wind power ultra-short-term prediction method in the field of offshore wind power plant cluster power prediction, and aims to solve the technical problems of spatial-temporal feature splitting and insufficient dynamic dependency relationship modeling. The method comprises the steps of performing anomaly detection and restoration on fan data, and generating a corrected wind power cluster data set; extracting a mean value, a standard deviation and a latest value of core operation data of each fan through a dynamic time window, and constructing a multi-dimensional node feature; a static geographic similarity matrix is generated based on geographic coordinates, a basic wake effect matrix is generated in combination with real-time wind direction data, correction is carried out through the maximum mutual information quantization time-delay effect, and then a dynamic adjacency matrix is obtained through self-adaptive fusion; and integrating the multi-dimensional node features and the dynamic adjacency matrix into a space-time diagram sequence data architecture, inputting the space-time diagram sequence data architecture into a multi-scale wake flow perception diagram space-time prediction model, and outputting a multi-fan power prediction value. According to the invention, high-precision multi-fan power prediction can be realized.

The invention provides a self-adaptive semantic-driven data set field matching method and system, and the system comprises a data preprocessing module which is used for carrying out the cleaning, standardization and preliminary analysis of an input data set, and extracting a field name, a data type, a field description and a data sample; the deep semantic representation modeling module is used for constructing a field-level semantic representation vector; the multi-level similarity calculation module is used for comprehensively calculating the grammatical similarity, the semantic similarity and the statistical similarity among the fields, dynamically adjusting the weight of the similarity of each level by adopting a weighted fusion algorithm, and generating a comprehensive similarity matrix; and the matching result management and application module is used for generating a field matching mapping table and a fusion suggestion according to the comprehensive similarity matrix. According to the method, high-precision automatic matching of data set fields is realized by fusing deep semantic understanding, multi-dimensional similarity calculation and incremental adaptive learning, and the efficiency and accuracy of data set fusion are remarkably improved.

The invention relates to technologies of federated learning, privacy protection and the like, belongs to the technical field of information security, and discloses a federated learning method for efficient client screening privacy protection for data isomerism, which mainly comprises two parts, namely a data sampling stage and a local training stage, generating a gradient similarity matrix by all the clients, and grouping the clients with high gradient similarity; in each group, the contribution value of each client to model training is calculated, the client with the high contribution value is selected as a training set, in the local training stage, a personalized privacy budget is generated according to the privacy requirements of the clients, and noise is injected into the local model according to the privacy budget to achieve differential privacy protection of model parameters. According to the method, the clients with high data quality can be screened out, the model convergence speed is increased, the personalized privacy requirements of the clients can be met, and relatively strong privacy protection is provided for the data.

The invention discloses a multi-level track disease recognition system based on vehicle body vibration data, and particularly relates to the field of blasting auxiliary devices. Comprising the steps that a data acquisition module acquires heterogeneity data of a track environment; the environment adaptive module analyzes the correlation between the vehicle body vibration data and the heterogeneity data, constructs a dual-view similarity matrix, and realizes adaptive classification of the track environment through a spectral clustering method; the label enhancement module generates high-quality pseudo labels and forms a completely labeled sample set; the disease recognition module adopts an attention-guided multi-task cascade CNN architecture, and outputs a multi-level recognition result containing disease positions and severity. By the adoption of the technical scheme, the problem that an existing track state evaluation method cannot evaluate the real state of track diseases in a complex environment is solved, and good performance is achieved in the aspect of geometric and structural disease recognition.

The invention provides a project cost big data management and analysis system, and relates to the technical field of data management, and the system comprises a data collection and preprocessing module which is used for collecting original cost data from a heterogeneous data source, and carrying out the preprocessing of the original cost data, and obtaining the preprocessed cost data; the semantic feature extraction module is used for converting the preprocessed cost data into a multi-dimensional feature vector based on a multi-level feature extraction system; the similarity calculation module is used for calculating similarities among different cost data based on the multi-dimensional feature vectors to obtain a similarity matrix; and the data storage and management module is used for storing the cost data, the multi-dimensional feature vector and the similarity matrix by adopting a mixed storage architecture, and providing retrieval and recommendation functions of cost projects based on a multi-level feature space index structure. According to the method, the limitation that a traditional method only depends on keyword matching is solved, and the system can recognize the deep incidence relation between the items.

The invention relates to the technical field of multi-view clustering, in particular to a multi-view clustering method and device, and can solve the problem that the overall effect of an existing method in a large-scale clustering task is limited due to the fact that the existing method has problems in the aspects of calculation efficiency, robustness and multi-view information integration to a certain extent. The method comprises the following steps: dynamically learning an anchor matrix and a projection matrix for each view, and constructing a bipartite graph to generate a similarity matrix; calculating a graph Laplacian matrix based on the similarity matrix of each view, and extracting spectrum embedding; the spectrums of multiple views are embedded and stacked into a third-order tensor, and cross-view shared information is extracted by using a low-rank tensor constraint; multi-view atlas embedding is aligned through a spectrum rotation technology, and a discrete clustering indication matrix is directly output.

The invention discloses a zero-trust architecture network security protection method and system, a medium and a program product, and the method comprises the steps: obtaining an access request; extracting credit characteristics, and calculating a user trust index according to the credit characteristics; constructing a directed trust transfer graph; calculating a trust attenuation coefficient of each trust transfer edge; according to the trust attenuation coefficient, calculating along the directed trust transfer graph to obtain a node trust value; constructing a relation similarity matrix; clustering according to the relationship similarity matrix to obtain a plurality of trust groups; calculating a group trust threshold according to the node trust value in each trust group; judging whether the node trust value of each trust node in the trust group in which the target resource is located is greater than a group trust threshold value corresponding to the trust group in which the target resource is located; if yes, safety verification is passed; and if not, updating the transmission path information and returning to execute the step of constructing the directed trust transfer graph according to the transmission path information. According to the invention, the security of the network is improved, and the network security protection effect is further improved.

A method for delineating an ecological corridor based on a maximum similarity model determines ecosystem type according to the determined ecological source, and assigns a habitat suitability index; and according to a maximum similarity matrix between different ecosystems, assigns the habitat suitability index to peripheral pixels of the selected ecological source, to obtain a similar value of the habitat suitability index. The habitat suitability index is corrected according to the surface curvature data and population density data, and resistance surface data is calculated according to the corrected value. Based on the resistance surface data, an ecological corridor is delineated comprehensively to find one or more paths corresponding to a minimum cost distance as ecological corridors. The influence of ecological environment on ecological corridor establishment is considered more comprehensively, so that the ecological corridor is delineated more accurately, and applied into practice to maximum extent.

The embodiment of the invention provides a video generation method and device and a storage medium, and the method comprises the steps: obtaining a video script, and obtaining a text feature of the video script; acquiring a video feature of each video material in the plurality of video materials; obtaining a similarity matrix between the text features and the video features; matching the video material with the video script based on the similarity matrix to obtain a video material sequence; and generating a target video file according to the video material sequence. According to the embodiment of the invention, through matching between the video script and the video material, retrieval and sorting of the video material are realized, so that video packaging is carried out based on the video material sequence, and automatic editing and production of the video are better realized.

The invention discloses a multi-channel short text matching method integrating external knowledge and syntactic structure, belonging to the field of deep learning technology, the method is based on the acquired text to be matched, based on the semantic relationship of the semantic original in the external knowledge base HowNet, the semantic similarity between words is calculated by topological distance, and the semantic original similarity matrix between words is determined; the attention weight of each word in the text is obtained by calculating the attention score of the text; according to the semantic original similarity matrix between words and the attention weight of each word in the text, the semantic feature vector of the text pair integrating the semantic original information is generated; by parsing the dependency relationship of the text and constructing a dependency structure graph, the nodes, edges and features of the graph are encoded as vector representations, and the syntactic structure feature vector of the text is obtained; the semantic feature vector and the syntactic structure feature vector are fused to obtain a comprehensive feature vector and input it into a classifier, and the text matching result is output. The method can achieve accurate matching of short texts.