21 results about "Cluster graph" patented technology

The invention relates to the technical field of bearing residual service life prediction, in particular to a bearing residual service life prediction method based on adaptive clustering graph convolution, and the method specifically comprises the steps: capturing vibration signal sample data from healthy to damaged of a bearing through a high-frequency vibration accelerometer; performing transformation processing on the collected vibration signal sample data to obtain time-frequency domain representation of sub-sample data; inputting the time-frequency domain representation of the sub-sample data into a bidirectional long-short-term memory network to obtain a time dependence feature; calculating two types of adjacent matrixes based on the time domain feature data, obtaining a final adaptive graph through control factor fusion, and introducing a node representation updating network to update node features in the adaptive graph so as to obtain spatial correlation information in the adaptive graph; and obtaining a final residual service life prediction result by using a clustering graph pooling method based on the current node features and the future node features. According to the invention, local information can be effectively aggregated, so that the prediction precision of the model is improved.

The invention provides a new energy power system frequency stability characteristic analysis method and device, equipment and a storage medium. The invention relates to the technical field of high-proportion new energy power system frequency stabilization. According to the method, a response characteristic curve based on key frequency characteristic quantity and frequency can be drawn in a new frequency characteristic distribution scene, firstly, a set of inertia center frequency and related characteristic quantity of a system is constructed, and after time sequence characteristic quantity is extracted, a thermodynamic diagram is drawn through Pearson correlation analysis; a hierarchical clustering graph is drawn through clustering analysis, so that the hierarchical clustering graph and the hierarchical clustering graph jointly present characteristic quantities with high correlation degree with the frequency; drawing a response characteristic curve of the system center frequency and the characteristic quantity with high correlation degree to form a criterion, wherein the response characteristic curve converges when the system frequency is stable; and the curve is divergent during instability. In addition, the influence of different new energy proportions on the frequency characteristic can be analyzed through the response characteristic curve.

Computer security improvements relating to fraud detection and data correlations through large-scale graph clustering of graph transformations and embeddings are disclosed. A service provider may utilize a framework having computing operations for detecting fraud and other malicious or suspicious activities by groups of accounts and fraudsters. In this regard, the service provider may transform relationship graphs of account networks and relationships between accounts and account data captured in the nodes and edges of such graphs. The service provider may merge nodes that edges connecting to other nodes of a certain type of account data, while other types of account data and nodes may not be merged. Edges may also be merged and weighted, and the resulting transformed graph may undergo graph embedding to generate vectors that may be clustered using an AI clustering algorithm. The clusters may then be used for AI model training and inferencing.

The invention discloses an unknown working condition equipment state anomaly detection method based on a graph neural network, and belongs to the field of mechanical equipment fault diagnosis. The method comprises the following steps: respectively preprocessing normal state data and abnormal state data of equipment with unknown working conditions, obtaining low-frequency data of different layers in each state, and determining an analysis layer; performing segmentation processing on data corresponding to the analysis layers in the two states; extracting features of each segment processing data from multiple angles; carrying out segmented processing on the data in each state to obtain a multi-modal feature node set, and generating a working condition graph feature set by adopting a feature centroid hierarchical clustering graph construction method; and obtaining a trained message passing graph neural network model by using the working condition graph feature set. According to the method, on the basis of constructing the working condition graph feature set, the anomaly detection of the equipment with the unknown working condition is effectively realized according to the constructed message passing graph neural network model.

The present disclosure relates to an approximate nearest neighbor search method and an approximate nearest neighbor search system. According to the embodiment, the approximate nearest neighbor search method manages graph index information for defining an inter-cluster graph. The approximate nearest neighbor search method searches for a vector closest to a query from vectors belonging to a search start cluster having a reference position closest to the query among a plurality of clusters. The approximate nearest neighbor search method selects one or more search target clusters close to the search start cluster while advancing along the inter-cluster graph, and searches for a vector closest to the query from vectors belonging to each of the one or more search target clusters. Thus, the approximate nearest neighbor search method capable of reducing the amount of data required to be rewritten accompanying update of the graph index and capable of obtaining sufficient search accuracy is provided.

The application discloses a tool wear resistance detection and characterization method. A wireless force measuring tool holder is used to clamp the tool, and cutting parameters are set to detect the tool wear resistance in real time. The tool wear resistance characteristic change graph is obtained, the angle change and numerical change range of each cutting edge wear resistance characteristic and the corresponding time are analyzed, and the wear resistance threshold is determined. Real-time data is processed by using a clustering model, and the clustering graph of the tool in different wear stages is compared with the initial state graph, so that the tool wear process is visually characterized, whether the tool wear resistance characteristic angle and value reach the threshold are judged, and whether the tool is seriously worn is determined. Combined with the clustering and regression methods, the characteristic data is recorded in real time, the position of the seriously worn cutting edge and the remaining use time are judged, and the tool wear resistance is determined. On the basis of clustering processing real-time data, regression calculation is further used, so that the calculation cost is reduced, and the problem that the specific cutting edge seriously worn position and wear degree cannot be accurately detected is solved.

The invention relates to the technical field of knowledge graph construction and information retrieval sorting, in particular to a technical requirement matching graph construction method based on a graph neural network, which comprises the following steps: reading technical requirement and technical supply literatures, extracting term entities and generating candidate pairs according to character string similarity and semantic similarity; constructing a candidate synonym cluster graph according to the candidate pairs, writing the candidate synonym cluster graph into a heterogeneous demand matching graph, and inputting the candidate synonym cluster graph into a Bayesian graph neural network to obtain node embedding and node and edge uncertainty parameters; weighting and aggregating the adjacency information based on the edge uncertainty and applying consistency constraint update in the cluster; and calculating a matching sorting score, writing a demand-supply matching edge according to a threshold value and Top-K, recording a confidence coefficient and an interpretation field, and generating and storing an atlas version. According to the method, the demand-supply matching accuracy is improved under the uncertainty constraint, and the mismatching rate is reduced.

According to one embodiment, an approximate nearest neighbor search method manages graph-based index information for defining an inter-cluster graph. The approximate nearest neighbor search method searches for a vector closest to a query vector from vectors belonging to a search start cluster that is closest to the query vector among a plurality of clusters. The approximate nearest neighbor search method selects one or more search target clusters close to the search start cluster while traversing the inter-cluster graph, and searches for a vector closest to the query vector from vectors belonging to each of the one or more search target clusters.

The invention discloses a tool durability detection and characterization method. A tool is clamped by adopting a wireless force measuring tool handle, cutting parameters are set, and the tool durability is detected on line in real time. And obtaining a tool durability characteristic change diagram, analyzing the angle change and value change range and corresponding time of the durability characteristic of each cutting edge, and determining a durability threshold value. And processing real-time data by adopting a clustering model, comparing clustering graphs of the cutter in different wear stages with initial state graphs of the cutter, visually representing the wear process of the cutter, judging whether cutter durability characteristic angles and values reach threshold values or not, and determining whether the cutter is seriously worn or not. And in combination with a clustering and regression method, characteristic data are recorded online in real time, the position of the seriously-worn cutting edge and the remaining use time are judged, and the tool durability is determined. And regression calculation is adopted on the basis of clustering processing of real-time data, so that the calculation cost is reduced, and the problems that the serious wear position of the specific cutting edge cannot be visually represented and the wear degree cannot be accurately detected are solved.

The application provides an event graph construction reasoning method and device based on sentence similarity. The method comprises the following steps: extracting a cause-effect relationship and event description from an original text, and constructing an initial event graph; calculating the similarity between nodes, i.e. event descriptions, in the initial event graph, merging events with a cause-effect relationship and a similarity greater than a set threshold into an event cluster to obtain an event cluster graph; inputting an event description into the event cluster graph, determining an event cluster in which an event with the greatest similarity to the event description is located, returning a partial event cluster graph having a direct cause-effect relationship with the event cluster, and calculating a related cause-effect event probability. The application merges events into event clusters based on similarity calculation, replaces events as nodes in a network structure, simplifies the network structure, and improves the efficiency of finding causes and results.

The application discloses an audio-visual cross-modal interaction design method for flat embroidery needle methods, which comprises the following steps: collecting different flat embroidery needle method graphs, vectorizing and sample augmenting each graph; using a K-means clustering algorithm to cluster graphs with similar features into clusters, screening representative graphs of each cluster to construct a visual modal information library; screening perceptual vocabulary capable of representing the visual modal information of the graphs, including attribute layer, perception layer and association layer vocabulary; scoring the matching degree of perceptual vocabulary at each level and each representative graph to construct the mapping relationship of "visual modal information-perceptual vocabulary" of each representative graph; finding out music bars related thereto, scoring the matching degree of each music bar and the representative graph, thereby constructing the "visual modal information-music bar" mapping relationship based on perceptual vocabulary; obtaining an audio-visual cross-modal information parameter table containing each representative graph and the corresponding music bar parameters based on the mapping relationship, adjusting the music bar parameters to output new music bars, evaluating and optimizing the matching degree of each representative graph and the corresponding new music bar, and outputting the optimized audio-visual cross-modal information library. The application produces a new interactive experience mode through scientific description of different flat embroidery needle methods to meet the needs of people for all-around perception and in-depth experience of embroidery technology.

The invention provides a text clustering method and device, electronic equipment, a storage medium and a program product, relates to the technical field of informatization software systems, and is used for improving the accuracy of text clustering. Inputting the abstract text into a first clustering model, and calculating a cosine similarity between the abstract text and a cluster center vector of each cluster in N clusters in the first clustering model through the first clustering model to obtain N similarity values; inputting the abstract text into a second clustering model, and calculating the matching degree between the abstract text and each cluster map in N cluster maps in the second clustering model through the second clustering model to obtain N first matching degree values; and determining a first cluster corresponding to the to-be-classified text based on the N similarity values and the N first matching degree values. The method is applied to a data classification scene.

The application discloses a news event prediction method based on heterogeneous evolution event clustering, comprising the following steps: generating event representation based on the preliminary updated entity representation and relationship representation in the constructed entity graph, regarding the event as a node, regarding the heterogeneous relationship between events as an edge, and constructing an event graph; obtaining event clusters by fuzzy clustering and constructing an event cluster graph; optimizing the event cluster representation according to the distance and similarity between event clusters on the event cluster graph by using a self-supervised optimization algorithm; capturing the implicit correlation between event clusters by using an implicit relationship encoder, and then updating the representation of the event cluster, the representation of the event, the entity and the relationship representation in sequence after sparsification and information aggregation; and predicting by a convolution-based news event model. The application effectively models the pair correlation, high-order correlation and multi-step time sequence evolution between events, and has important application value in international situation analysis, social governance and intelligent decision support.

The application discloses a news event prediction method based on time smoothing constraint deep evolution clustering, comprising the following steps: constructing an entity graph of a news event; using a relation-aware graph convolutional neural network to aggregate information of the entity graph; obtaining a soft membership matrix of the entity to the cluster by using fuzzy clustering in a deep evolution clustering module, and constructing a cluster graph by aligning and fusing clusters of adjacent time stamps; using implicit correlation between cluster pairs in a cluster graph information transmission module to aggregate information to update the representation of the cluster and the representation of the entity and the representation of the relation; combining the updated representation of the entity and the representation of the relation with the initial representation of the entity and the representation of the relation as next time stamp input by using a time residual gate; inputting the updated representation of the entity and the representation of the relation of different time stamps into a time-dependent encoder based on an attention mechanism to obtain comprehensive representation of the entity and the relation; and predicting by a news event model based on convolution. The application can effectively model the time sequence evolution of high-order correlation between entities.

The invention provides an information security risk detection method and system based on data feature analysis, and the method comprises the steps: obtaining traffic data of each node of a network, extracting communication protocol state transition and load information entropy features, and constructing a node feature vector through combining port and protocol information; clustering the nodes by adopting a density peak clustering algorithm to obtain a plurality of initial clusters; an initial cluster is regarded as a super node, a communication stability coefficient is determined according to communication flow jitter and periodicity between the super nodes, a cluster graph is constructed, and node feature vector dispersion in the super nodes is calculated to serve as an internal stability index; calculating super-node structure centrality based on the cluster graph, fusing the node feature vector and the subordinate super-node structure centrality to obtain a node context risk value, and aggregating node risk values in the super-nodes to generate a cluster risk value; and determining a global risk baseline according to statistical distribution of all cluster risk values, adjusting the baseline in combination with a super-node internal stability index to obtain a threshold value, and determining that the cluster is a high-risk cluster if the cluster risk value exceeds the threshold value.

The application discloses a kind of automatic driving multi-vehicle coordination control method and system based on robust QMIX, belong to vehicle networking and automatic driving field, including: by cluster graph representation module, the local graph structure of description in cluster vehicle interaction and the global graph structure of inter-cluster interaction, to represent the interaction relationship inside and outside vehicle cluster;Through the noise perception mechanism of robust strategy learning module, dynamically evaluate the environmental disturbance intensity, adopt the weighted QMIX framework to adjust the value function weight, and combine the robust loss function to optimize strategy;Through cross-cluster reward coordination module, according to the reward distribution difference between clusters, dynamically reduce the reward value of high-reward cluster to balance global optimization;Through central dispatching system, cooperative decision-making instruction is distributed to each automatic driving vehicle for execution.The application can realize long-term stable strategy optimization in complex traffic system.

The invention provides a patient primary index management method and device, and relates to the technical field of computers. The method comprises the following steps: acquiring field data of target fields in a plurality of candidate records; determining a matching decision factor of the record pair, and identifying the record pair as a suspected matching pair when the matching decision factor falls in the suspected matching interval; constructing at least one suspected cluster graph model based on the matching decision factor of the suspected matching pair; iteratively optimizing the edge selection in each suspected cluster graph model until the consistency in the suspected cluster graph model is maximized and the conflict edge selection is minimized, and dividing the suspected cluster graph model into a first suspected cluster to be merged and a second suspected cluster to be audited based on the merging tendency strength of the current edge; performing depth feature analysis on the second suspected cluster by using a depth discrimination model to obtain an auxiliary decision label; and updating the patient main index database according to an audit confirmation result from the user terminal. The processing efficiency of patient primary index management is improved.

System and apparatus implementing linear error correction codes and associated linear error correcting codes are provided. The apparatus includes a receiver for receiving a sequence of symbols via a channel, the sequence of symbols having been encoded at a source using a linear error correcting code. The apparatus includes a decoder for decoding the received sequence of symbols and outputting a decoded sequence of bits. The decoder is configured based on a multivariate message-based graph representation (e.g. cluster graph or region graph) of the linear error correcting code. The linear error correcting code (low-density parity check code (LDPC) or a polar code) may be configured such that a metric (e.g. average factor overlap) based on an overlap of one or more parity checks factors with one or more other parity check factors is greater than 1. A method may also comprise (Figure 1C) receiving a sequence of bits encoded at a source using a linear error correcting code, initialising parity check clusters of a cluster graph representation, initiating message passing between the parity check cluster and when a consensus between all the parity check cluster regarding shared bit values is reached, outputting a decoded sequence of bits.

The present application relates to the technical field of knowledge graph construction and information retrieval sorting, and particularly relates to a technical requirement matching graph construction method based on a graph neural network, comprising: reading technical requirement and technical supply literature, extracting term entities and generating candidate pairs according to string similarity and semantic similarity; constructing a candidate synonym cluster graph according to the candidate pairs and writing into a heterogeneous requirement matching graph, inputting a Bayesian graph neural network to obtain node embedding and node and edge uncertainty parameters; weighting and aggregating adjacent information based on edge uncertainty and updating within the cluster by applying consistency constraints; calculating a matching sorting score, writing requirement-supply matching edges according to a threshold and Top-K, recording confidence and explanation fields, and generating and storing a graph version. The present application improves requirement-supply matching accuracy under uncertainty constraints and reduces the mis-matching rate.

This invention relates to a method for predicting urban traffic flow based on Markov clustering graph attention networks, comprising the following steps: 1. Obtaining a time-series traffic flow matrix based on historical traffic data; 2. Extracting natural structural information from the graph based on the Markov clustering algorithm to obtain a global relevance node matrix; 3. Establishing a generative adversarial neural network model, wherein the improved graph attention module in the generator, when acquiring spatial hidden features, no longer restricts neighboring nodes to only first-order neighboring nodes as in graph attention networks, but extends to the global relevance node information obtained based on the Markov clustering algorithm; learning and training the model, and using the learned model as a regional traffic flow prediction model; the improved graph attention module of this invention not only focuses on local neighboring nodes, but also dynamically considers the neighboring node information in the overall graph structure, assigning different weights to them, thereby improving the ability to acquire spatial features.