62 results about "Pattern clustering" patented technology

The invention discloses a black-box test zero-day vulnerability analysis method and system based on multi-dimensional data, and aims to solve the problems that a traditional black-box test means is weak in unknown vulnerability recognition capability, high in false alarm rate, lack of path modeling and verification mechanisms and the like. The method comprises the following steps: constructing a cross-time window behavior graph by acquiring multi-dimensional heterogeneous information such as network input, system call, log information and abnormal signals; on the basis, potential abnormal paths are identified through structure entropy change and graph structure mutation analysis, path vector representation is constructed by combining graph representation learning and a path embedding method, and path-level risk modeling and mode clustering analysis are achieved; furthermore, vulnerability confirmation is carried out on the suspicious path through multiple verification mechanisms such as attack replay, fuzzy testing and sensitive function combination identification. The system has a multi-module cooperation capability, can realize automatic mining, verification and visual tracing of zero-day vulnerabilities in a source-source-free environment, and has good universality and expansibility.

The invention relates to an abnormity alarm method and system for a feed production control system, and the method comprises the following steps: collecting multi-field data related in a feed production process in real time, the multi-field data at least comprising production equipment operation parameters, material characteristic parameters, environmental parameters and process control parameters; preprocessing the collected multi-field data, and fusing to generate a multi-field description feature vector representing the current production state; inputting the multi-field description feature vector into a pre-trained fault prediction model, wherein the fault prediction model respectively calculates the confidence of the fault type prediction result and the confidence of the fault point prediction result; wherein the confidence coefficient is in positive correlation with the distance from the feature vector to a known fault mode clustering center or the distance from the feature vector to a decision boundary; and when the fault prediction result indicates that an abnormal risk exists and the corresponding confidence exceeds a preset threshold, triggering a corresponding alarm signal.

The invention discloses a CKD special disease database construction method and system, and relates to the technical field of database construction, and the method comprises the steps: obtaining and predicating a multi-source heterogeneous atomic fact; constructing and fusing three sets of ontology models; verifying and deducing facts in a layered manner according to a multi-layer priority rule base; calling a constraint completion engine to generate inference facts for missing or conflicting fields; streaming follow-up visit, test and intervention logs into events and triggering graded alarms; an anti-fact intervention scene is simulated for the high-risk patient, and pilot groups are screened; dynamically evaluating and adjusting the rule priority; and based on a causal chain and an event mode, clustering patients and constructing a generality map, and recording a whole-process operation chain. Semantic management is realized through fusion of ontology, anti-fact intervention simulation is performed on high-risk patients by means of multi-layer rule verification, and database semantic consistency, reasoning depth, decision support and traceability are improved through causal atlas driven clustering.

A method of clustering patterns of an integrated circuit includes; providing a pattern image and numeric data, as input data corresponding to a first pattern to a first model, wherein the first model is trained by a plurality of sample images and a plurality of sample values, obtaining a content latent variable using the first model, and grouping a plurality of content latent variables corresponding to a plurality of patterns into a plurality of clusters based on a Euclidean distance, wherein the numeric data represents at least one attribute of the first pattern.

The invention provides a charging and discharging strategy intelligent optimization method and system applied to an energy storage power supply, and the method comprises the steps: carrying out the time sequence segmentation processing of a historical charging and discharging process of the energy storage power supply, employing a sliding time window to traverse historical charging and discharging data, and generating a charging and discharging time period data sequence with a continuous time stamp; performing charging and discharging feature extraction on the charging and discharging time period data sequence to obtain a charging and discharging feature set corresponding to each charging and discharging time period, performing mode clustering analysis based on the charging and discharging feature sets, dividing the charging and discharging time periods with similar feature distribution into a plurality of charging and discharging mode clusters through time sequence clustering, and performing mode clustering analysis on the charging and discharging mode clusters; and carrying out association matching on the charging and discharging mode clusters and power grid load prediction information, carrying out strategy iteration optimization processing on an association matching result through reinforcement learning, and generating a charging and discharging power distribution scheme adapted to different charging and discharging mode clusters. According to the invention, the optimization accuracy and environmental adaptability of the charging and discharging strategy of the energy storage power supply can be improved.

The invention relates to the technical field of electrochemical energy storage systems, and discloses an electrochemical energy storage system battery charging effect evaluation method, which comprises the following steps: acquiring energy storage system operation data and battery charging records, extracting time sequence characteristics, equipment state characteristics and charging parameter characteristics, and calculating the charging effect of an electrochemical energy storage system battery. Constructing a multi-dimensional feature vector and generating an initial charging effect feature library; the evaluation model is dynamically adjusted based on feature library feedback, potential effect features are identified through mode clustering analysis, feature priorities are optimized, and model parameters are updated; and in combination with system configuration information and target model parameters, determining an optimal power compensation effect evaluation result by adopting a hierarchical disassembly and classification statistical method. According to the method, through multi-dimensional feature fusion, dynamic model optimization and refined statistical analysis, the accuracy and adaptability of electricity supplement effect evaluation are remarkably improved, and a scientific basis is provided for efficient operation and maintenance of the electrochemical energy storage system.

The invention discloses a plum rain season photovoltaic output prediction method and system based on small sample learning. The method comprises the following steps: constructing a multi-source feature system; clustering and dividing meta-learning tasks based on a meteorological mode; the method comprises the following steps: constructing a TCN-Transform hybrid neural network model, and embedding a learnable meteorological adapter; training a hybrid neural network model by adopting an MAML meta-learning framework, and realizing meta-learning task training of different meteorological models in combination with a physical constraint regularization loss function; model parameters are optimized in three stages through a course learning strategy, and a training set is enhanced by utilizing Mixup time sequence data; and deploying the trained model to carry out output prediction, and outputting a prediction result with a confidence interval. According to the method, high-precision photovoltaic output prediction in a small-sample and high-uncertainty plum rain season scene is realized, and reliable support is provided for power grid dispatching.

The invention provides a log data-oriented system fault mining method, which belongs to the technical field of information, and comprises the following steps: collecting original log data of a system, and preprocessing original system logs, including analysis, key information extraction, vectorization and generation of a structured log event sequence; processing the log event sequence by adopting a two-stage model fusing anomaly detection and mode clustering so as to distinguish accidental anomaly and a real fault mode; and for each candidate fault mode cluster, analyzing a time sequence causal relationship between internal events, and positioning a root cause event according to the causal flow score of the node. The method has the advantages that dependence on artificial experience is reduced through full-process automatic modeling and analysis; according to the method, log analysis, feature extraction, anomaly detection and root cause positioning are all automatically completed by adopting a preset algorithm process, and the detection effect is maintained through self-adjustment of the model, so that the flexibility and sustainability of fault mining are improved.

The invention relates to the technical field of data processing and analysis, in particular to a knowledge graph entity accurate identification relation system for a service system, which comprises the following steps of: analyzing database table data, extracting an attribute change track and generating an attribute evolution sequence to provide a time sequence data basis for subsequent analysis and ensure the traceability of attribute evolution; meanwhile, the adaptability of the system to heterogeneous data is enhanced; clustering and dividing attribute clusters based on an attribute change mode, establishing a cross-cluster association index, revealing a potential association rule between attributes, and supporting rapid classification and dynamic adjustment of attribute relationships in a service scene; dynamic analysis of intra-cluster and cross-cluster attribute influence is realized by using a bidirectional attribute propagation model, and coupling strength between attributes is accurately quantified through iterative calculation and feedback calibration; by constructing a dynamic attribute topological graph and monitoring attribute value changes in real time, association reconstruction is triggered to generate an attribute association group set, dynamic changes of a service system are adapted, and association accuracy is improved.

The invention relates to the technical field of assisted reproduction health management, and discloses an assisted reproduction management method and system. The method comprises the steps that physiological data of multi-source monitoring equipment is received and subjected to standardization processing, and a multi-dimensional feature matrix is formed through time sequence alignment and feature fusion; identifying a dynamic evolution track of the key physiological mode by utilizing a mode clustering and sliding window technology; performing state deduction on the mode sequence based on the personalized model, and predicting a future physiological state; generating a regulation and control strategy sequence according to the difference between the prediction and the target, and executing the instruction; physiological response data is collected in real time and compared with an expected model, a follow-up strategy is dynamically adjusted, and closed-loop optimization is formed; and finally evaluating path validity and updating model parameters according to complete cycle data. According to the method, continuous accurate prediction and individualized self-adaptive regulation and control of the reproductive physiological state are realized, and the management accuracy and efficiency are improved.

The application provides a resource scheduling method and system based on campus user behavior analysis, and relates to the technical field of resource management. The method comprises the following steps: collecting target behavior information of a target campus user in a preset period, and analyzing a target behavior time sequence; acquiring an arbitrary time zone and matching arbitrary behavior data; forming a campus user set, and analyzing first behavior information of a first user; if a first mode index of the first behavior time sequence and the target behavior time sequence reaches a predetermined index limit value, a mode clustering cluster is formed; acquiring a number of user individuals in the mode clustering cluster, and combining the arbitrary behavior data to obtain arbitrary resource demand of the arbitrary time zone; and performing campus resource scheduling processing. The application solves the technical problem that, in the prior art, inaccurate prediction of user demand leads to unreasonable resource allocation, thereby affecting resource scheduling efficiency. By analyzing user behavior mode and determining resource demand, the efficiency of resource scheduling is improved.

The application discloses a fine-grained load segmentation prediction method, and the steps are as follows: based on the daily load curve, dimension reduction indexes are extracted, and a user power consumption mode clustering model is established; a multi-task group load prediction model is constructed; different categories of user daily load curve segmentation is completed; the fluctuation rate, fluctuation standard deviation, fluctuation degree, high energy consumption proportion and low energy consumption proportion indexes of different fluctuation segments are analyzed, and the type of the fluctuation segment is identified; a group prediction result direct correction prediction model is established for the rising and falling trend load fluctuation segment; a cyclic classification correction prediction model is established for the high energy consumption fluctuation, low energy consumption fluctuation, sharp peak fluctuation and oscillation fluctuation segment; and a fine-grained user complete load curve prediction model is constructed. The application is aimed at fine-grained load which is difficult to predict, improves the prediction accuracy, ensures that the power grid can accurately master the fine-grained load curve change situation, matches the power demand situation of the end user at any time, and helps the rapid development of intelligent energy consumption and point-to-point power transaction.

The invention relates to the technical field of titanium alloy smelting analysis, in particular to a titanium alloy smelting effect intelligent analysis method and system, and the method comprises the steps: recording a smelting event of smelting equipment in a preset time period; recording abnormal events of smelting events in each time of smelting in a single parameter and parameter pair combination mode to form an associated parameter chain; performing mode clustering on the associated parameter chain according to a combined single parameter and parameter pair form, and determining a propagation path after clustering; based on the propagation path, performing single event scoring on each batch of abnormal events, and constructing a batch-index incidence matrix; according to the change trend of each key index in the batch-index incidence matrix during multiple times of smelting, obtaining the transfer relation of abnormal events in multiple times of smelting; determining target parameters of each batch on the basis of a passing relation of abnormal events in multiple smelting, and outputting the target parameters; the accuracy of smelting analysis and the efficiency of dynamic perception are realized.

The invention provides a multi-type railway track inspection data mining and state evaluation method based on knowledge embedding and cross attention. The method aims at solving the key technical problems of adaptive feature extraction, high-dimensional spatial pattern clustering, online real-time evaluation and the like of multi-source heterogeneous data. By constructing a technical system fusing domain knowledge embedding, multi-modal feature alignment and dynamic model optimization, full-process intelligentization from data preprocessing, feature extraction to state evaluation is realized, and an efficient and accurate solution is provided for intelligent operation and maintenance of rail transit.

The invention discloses an enterprise data intelligent storage optimization method and system. The method comprises the steps of enterprise data collection, data access mode preliminary screening, data access mode clustering, alternating refinement and enterprise data storage optimization. The invention belongs to the field of data processing, and particularly relates to an enterprise data intelligent storage optimization method and system.According to the scheme, a dispersion popularity membership degree is introduced, a weakening coefficient is added into a target function, historical access influences are attenuated, the weight of burst high-frequency access is reduced, and then misclassification caused by accidental access is reduced; business association driven local self-adaptive approaching clustering is performed on enterprise data, so that a high-benefit storage scheme can be formulated in a unified manner; a common precision objective function is constructed, and bidirectional matching of the mode and enterprise data association degree is ensured through an alternate precision framework; and the enterprise data intelligent storage optimization effect is improved.

This invention relates to an intelligent deformation discrimination method based on InSAR temporal deformation patterns and building semantic information, comprising: acquiring InSAR temporal monitoring data and building vector data of the target monitoring area; performing temporal deformation pattern clustering on discrete deformation scattering points to determine the deformation pattern category label of each discrete deformation scattering point; identifying discrete deformation scattering points falling within the polygonal area of ​​the building outline as the internal point set of the building, and identifying discrete deformation scattering points falling within the surrounding foundation ring zone as the surrounding foundation point set; extracting a building-level comprehensive feature vector, including basic statistical feature components, temporal pattern distribution features, and internal and external difference features, using a single building as a unit; inputting the building-level comprehensive feature vector into an unsupervised anomaly detection model to filter out anomaly building sets; inputting the features of the anomaly building sets into an anomaly attribution clustering model, and determining the anomaly cause category of each anomaly building based on the clustering results.

This invention discloses a method and system for battery aging assessment of battery swapping cabinets based on charging pattern clustering, belonging to the field of battery testing technology for battery swapping cabinets. The method includes: analyzing battery charging cycles according to the charging cycle of the swapping cabinet to construct multi-dimensional charging features; performing clustering learning based on these features to construct multiple charging pattern clusters; performing health analysis on the battery based on these clusters to obtain battery health status parameters; retrieving real-time charging data and matching the charging pattern clusters with the health status parameters to determine the target charging pattern cluster; and performing battery aging assessment based on the target charging pattern cluster to generate aging assessment results. This invention solves the technical problems of low assessment accuracy and poor real-time performance caused by the diversity of charging patterns in existing battery aging assessment methods. It achieves refined analysis and dynamic matching of battery health status based on charging pattern clustering, improving the accuracy, scenario adaptability, and real-time responsiveness of battery aging assessment results.

The present application provides a method and system for evaluating the moisture permeability of concrete. The method first obtains moisture gradient change data within the concrete and performs moving average processing on the data to generate smoothed moisture diffusion data. Secondly, based on the diffusion data, the frequency offset data caused by humidity influence is measured in the microwave frequency band, and diffusion features are extracted from the offset data using phase locking technology. The diffusion features are then subjected to spatiotemporal correlation processing to identify hotspot patterns. Finally, pattern clustering processing is performed on the hotspot patterns, grouping them into multiple levels and outputting the concrete permeability rating. The technical solution provided by the present application not only ensures real-time and lossless capture of the microscopic dynamics of moisture migration, but also achieves zero-delay analysis of the true permeability dynamics, fundamentally resolving problems such as manual experience misjudgment and offline simulation inaccuracy.

The invention relates to the technical field of power system frequency modulation capacity demand prediction, and provides a frequency modulation capacity demand prediction method considering fluctuation mode clustering. The method comprises the following steps: collecting historical operation data of a regional power grid to be predicted to obtain system net load data; dividing a frequency modulation capacity demand mode for the system net load data by using K-means clustering considering local neighborhood information; according to the different frequency modulation capacity demand modes, using an optimization-based random forest feature selection method to screen an optimal prediction feature subset; and predicting the frequency modulation capacity demand by using a bidirectional long-short term memory neural network probability model according to the optimal prediction feature subset. According to the method, the improved K-means clustering, the optimized random forest feature selection method and the bidirectional long-short term memory neural network probability model are fused, so that the rapid frequency response of the power grid is realized, the economical efficiency of the system is improved, and the safety degree of operation is improved.

The invention relates to the technical field of cloud data processing, and discloses a cloud financial data analysis and early warning device which comprises a cross-domain event aggregation module, a dynamic graph construction module, a causal flow inference module, a pattern convergence analysis module and an interactive presentation module. The method comprises the following steps: aggregating business events from a plurality of business systems, and constructing a dynamic business process map; then, identifying an abnormal disturbance event in the map, and generating one or more abnormal causal chains connected to the quantitative financial influence based on the event tracing so as to generate tactical early warning; the device performs pattern clustering and root cause analysis on the plurality of abnormal causal chains to generate strategic insights. According to the method, the dynamic graph is constructed, and causal inference and mode convergence are executed, so that the problems that traditional financial analysis is lagged and a deep causal relationship between business and finance is difficult to reveal are solved, and prospective and penetrating analysis and early warning from a single business event to a systematic risk root cause are realized.

The invention relates to the field of multi-section lightning arrester state monitoring, in particular to a multi-section lightning arrester state synchronous monitoring system. According to the system, firstly, historical operation state vectors formed by all types of historical operation data of all node lightning arresters at each time point are clustered, and a plurality of normal mode clusters are obtained; according to the distance between a real-time operation state vector formed by all types of real-time operation data of each node lightning arrester at the current time point and the clustering center of the normal mode clustering cluster, the operation deviation degree of each node lightning arrester at the current time point is obtained; and according to the operation deviation degree of each node lightning arrester and the adjacent node lightning arrester at the current time point and the distribution of the historical operation state vectors in the normal mode cluster, carrying out whole group fault early warning on the multi-section lightning arrester group. The method can adapt to the dynamic change of cooperative operation of multiple sections of lightning arresters, so that the fault condition of the whole group can be accurately monitored.

The application discloses a self-adaptive building day-ahead load prediction method based on transfer learning and relates to the technical field of buildings and environmental protection.The application comprises the following steps: S1, data acquisition and processing, wherein original data sets are divided into small data sets of target buildings and large data sets of basic building groups, and missing values of all original data sets are filled; S2, use mode clustering; S3, source domain data screening, historical daily load curves of use modes of load target buildings are screened, and a data transfer training set and a model transfer training set are respectively constructed; S4, day-ahead load prediction model construction; and S5, self-adaptive model optimization, wherein model parameters are continuously adjusted by using Bayesian optimization, and self-adaptive load prediction of target buildings is realized.The application realizes load prediction of target buildings by combining historical data of data-sufficient building groups with the data transfer and model transfer methods of transfer learning.

The invention discloses a supply chain demand prediction method based on pattern clustering, and the method comprises the following steps: S1, collecting multi-source data, and constructing a multi-source original data set; s2, preprocessing the original data set to generate a standard data set; s3, based on improved HDBSCAN clustering analysis, identifying a potential demand trend and dividing an output demand mode group; s4, extracting a key feature vector of each group, inputting the key feature vector into a DeepAR model for autoregressive training and cross validation, and outputting a prediction model; s5, loading the trained model, and combining real-time sales and market data to generate a current key feature vector; and S6, inputting the current feature vector into a DeepAR model, predicting a future demand, and outputting the future demand to a supply chain scheduling system. According to the invention, automatic identification and high-precision prediction of the demand trend under the driving of multi-source data are realized, and the supply chain response, efficiency and inventory decision intelligent level are improved.

Systems and methods for simultaneously analyzing time-series and dominant frequency data from both a process domain and an electrical domain of a facility, such as an industrial plant, to detect instances of deviation from optimal, normal, or other predefined process conditions or electrical trips. Detecting when changes in frequencies occur in the time-series data creates a time-series of the changes in dominant frequencies. A data analysis server detects instances of deviation from the predefined process conditions or electrical trips by detecting one or more of correlations, patterns, clusters, and the like in the rates of change. The data analysis server employs one or more of statistical analyses, data mining, machine learning, deep neural networks, parallel coordinate analyses, etc. to identify the deviations and predict or detect onset of an undesired event such as a process perturbation or electrical trip.

The invention provides a commercial building energy consumption anomaly detection method based on a graph method, and relates to the technical field of intelligent constructions.The method comprises the steps that a graph structure is constructed based on similarity, nodes are mapped into sub-item and sub-partition sub-tables, and edge weights represent time sequence similarity between the corresponding nodes; generating spectral distribution representing topological characteristics of the graph structures by calculating a Laplacian matrix eigenvalue set of the graph structures, and performing hierarchical clustering on the plurality of graph structures based on a Batta-just distance between the spectral distribution to obtain energy consumption mode clustering results in different time periods; and carrying out contour coefficient analysis on the clustering result, screening a class cluster number corresponding to a maximum contour coefficient value as a corresponding partition, and defining a class cluster set containing historical normal data as a baseline class cluster. According to the invention, the accuracy of building energy consumption anomaly detection can be improved.

The application discloses an engine group fault mode recognition method and device based on density clustering-support vector machine and multi-time point prediction weighting and electronic equipment, and the method comprises the following steps: performing dimension reduction processing on training engine performance parameters and test engine performance parameters respectively based on principal component analysis (PCA), to obtain training engine low-dimensional performance parameters and test engine low-dimensional performance parameters; in the prediction of engine single-time point fault mode, performing fault mode clustering processing on the training engine low-dimensional performance parameters based on a density-based spatial clustering of applications with noise (DBSCAN) algorithm, to obtain corresponding fault mode classification labels of each training engine; a single-time point fault mode classification model based on a support vector machine (SVM) is established, and the single-time point fault mode classification model based on the SVM is trained by using the training engine low-dimensional performance parameters and the corresponding fault mode classification labels of each training engine, to obtain a trained single-time point fault mode classification model based on the SVM; the test engine low-dimensional performance parameters are input into the trained single-time point fault mode classification model based on the SVM, to obtain single-time point fault mode classification labels of each test engine; and finally, based on multi-time point fault mode weighted prediction and the single-time point fault mode classification labels of each test engine, final fault mode classification labels of each test engine are obtained.

The invention relates to a file box optimization design printing method based on an electronic file system, and relates to the field of data processing, and the method comprises the steps: downloading an electronic file set with a file box identifier, executing file box shape and size clustering, and obtaining a primary file clustering result; performing file box cover pattern clustering on the primary file clustering result to obtain a secondary file clustering result; performing cover pattern layout position clustering on the second-level archive clustering result to obtain a third-level archive clustering result; after the electronic archive system is initialized according to the three-level archive clustering result, a to-be-packaged archive image uploaded by a user side is received, the electronic archive system is traversed for archive matching, and matched archive categories are obtained; and when the matched archives belong to the same class, designing and printing the archive box according to the class archive box identifier, so that the technical problems of low actual production efficiency and insufficient systematic management caused by incapability of realizing automatic customized printing of the archive box are solved.

The invention provides a CFRP reinforced ancient building wood structure flexural capacity calculation method based on finite element analysis, relates to the technical field of model simulation, and aims at providing a calculation framework for converting historical text big data into finite element analysis prior guide information. According to the framework, qualitative historical vulnerability evaluation and quantitative physical and mechanical simulation are seamlessly integrated through an innovative probability distribution weighting mechanism. Simulation sampling is guided through historical information, the number of finite element analysis times needed for achieving the same confidence degree is reduced, and the calculation cost and time are reduced. The simulation process focuses on a weak link which is more likely to exist in history, the recognition capability of a real and specific risk mode of the structure is effectively improved, and the misjudgment rate caused by disconnection of model hypothesis and historical reality is reduced. Finally, a reinforcement intervention priority list fusing failure mode clustering analysis and component importance evaluation is output, and direct, quantitative and data-driven decision support is provided for engineers.

The invention discloses an enhanced fingerprint-based DDoS attack recognition, pattern clustering and tracing method and system, and the method comprises the steps: extracting multi-dimensional features, including basic features, application layer behavior features, load entropy features and time sequence statistical features, from network traffic, and calculating derivative features, including time sequence features and rate features; carrying out normalization processing on the multi-dimensional features and the derivative features, and carrying out weighted fusion to form an enhanced multi-dimensional traffic fingerprint vector; performing DDoS attack traffic preliminary screening on the enhanced multi-dimensional traffic fingerprint vector through a dual-threshold preliminary screening logic, performing secondary verification through a trained Bi-LSTM, and performing classification; clustering is carried out on the identified DDoS attack traffic; and tracing the identified DDoS attack traffic according to the network topology information and the traffic log. According to the method and the system, accurate identification, classification and traceability of the DDoS attack are realized.

A method of evaluating an integrated circuit, a method of emulating an integrated circuit, and a system of evaluating an integrated circuit are provided. The method of evaluating an integrated circuit includes: obtaining a plurality of patterns representing a layout of the integrated circuit; clustering the plurality of patterns into a plurality of clusters based on geometric features of the plurality of patterns and simulation results obtained by simulating properties of the plurality of patterns; selecting a representative pattern for each of at least one cluster of the plurality of clusters; and validating the representative pattern for each of the at least one cluster and evaluating performance of the integrated circuit based thereon.