24results about How to "Accurate clustering" patented technology

The invention discloses an environment monitoring system and method for a modern poultry farm, and relates to the technical field of data processing. The environment monitoring method for the modern poultry farm comprises the following steps: acquiring a temperature change sequence of each breeding area in the poultry farm, wherein the temperature change sequence is used for representing the temperature change condition of the breeding area in a preset time period; by means of the temperature change sequences, the clustering coefficient of regional feature points in the breeding regions at the current moment is determined, and the regional feature points are preset feature position points in the breeding regions; according to the clustering coefficient of each regional feature point at the current moment and the distance between the regional feature points, clustering each regional feature point to obtain at least one current cluster region; determining the poultry gathering rate of the current cluster region by using the mass center position change information and the region range change information of the current cluster region; and determining whether the environment temperature of the poultry farm is abnormal or not according to the poultry gathering rate of the current cluster region.

The invention discloses an on-demand content clustering method, an on-demand content clustering device, electronic equipment and a storage medium, aiming at data characteristics in an IPTV (Internet Protocol Television) service environment, and optimizing a media resource clustering algorithm cooperating with user behaviors. And in combination with the media asset content and the user data, the clustering accuracy can be effectively improved. In addition, large model feedback information is innovatively introduced to correct the IPTV recommendation model. By constructing a two-tuple task and inquiring whether the two large models are similar or not, the weight or parameter of the recommendation model can be adjusted according to the feedback of the large models, so that the clustering effect is further optimized. The IPTV multi-dimensional fine granularity classification method is remarkably innovated and improved, has the advantages and effects of efficient resource utilization, multi-dimensional classification, vectorization strategy innovation, efficient model training and optimization, promotion of business innovation development and the like, and can be widely applied to the technical field of data processing.

The present invention relates to medical image processing technology, and discloses an infectious disease risk assessment method and system based on multimodal graph coding. The infectious disease risk assessment method based on multimodal graph coding is applied to a PACS system, and the method includes: extracting image features from input image data, and determining the image sample category based on the image features; splicing text features with image features, reducing the dimension, and obtaining the spliced ​​graph code through a two-layer graph neural network output, and inputting the spliced ​​graph code into a kmeans clustering model to determine the category of the inspection sample; after clustering, inputting the number of newly added samples of each category on the same day into an early warning model to determine its confidence; comparing the obtained confidence with a set threshold value to determine its risk coefficient. The present invention uses a graph neural network to convert simple pattern recognition features into graph codes, emphasizing the connection between similar samples, making the kmeans clustering effect more accurate, and thus more sensitive to risk assessment.

The invention discloses a heterogeneous relation convolution spectral clustering method for incomplete multi-view streaming news data, and belongs to the field of data clustering processing in the incomplete multi-view streaming news data. And constructing a heterogeneous relation matrix and a high-dimensional manifold distribution matrix for each view. And then inputting the sample features of each view and the corresponding heterogeneous relationship matrix into an incomplete multi-view streaming news data heterogeneous relationship convolutional spectral clustering network to obtain a low-dimensional embedded representation of each view, and obtaining a common low-dimensional embedded representation of all views through a feature fusion layer. Then, the common low-dimensional embedding representation is input into an orthogonal normalization layer to obtain approximate spectrum embedding, and then the approximate spectrum embedding is processed through a self-supervision module to obtain soft distribution of single input sample features. And finally, optimizing the heterogeneous relation convolution spectral clustering total loss function of the incomplete multi-view streaming news data to convergence by using a small-batch stochastic gradient descent algorithm. And when the model reaches convergence, clustering single incomplete multi-view streaming news data samples continuously generated after the t0 moment by using the model to obtain a clustering result of single streaming news data, thereby realizing the purpose of clustering the incomplete multi-view streaming news data.

The invention relates to a model training method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the steps of receiving multiple groups of first model parameters fed back by multiple clients, determining the first model parameters as sample points to obtain a sample data set, determining distance inverse number density of the sample points for each sample point based on the distance between the sample point and adjacent sample points of the sample point, and determining the distance inverse number density of the sample points based on the distance inverse number density. The method comprises the following steps: determining a local density correction term of a sample point, clustering a sample data set based on the distance inverse number density and the local density correction term, determining the local density of the sample point, and based on the local density of each sample point, dividing clients with similar data distribution into the same client set to obtain a plurality of client sets, and scheduling the plurality of clients. Clients in the same client set share model parameters, and the model is jointly trained in a federal learning mode. By adopting the method, the model training precision can be improved.

Some embodiments of the present application provide a method, device, electronic device and storage medium for clustering target objects, the method comprising: obtaining multiple target objects; projecting each target object into a grid template respectively, and extracting a feature sequence corresponding to each target object; and obtaining clustering results for the multiple target objects according to the feature sequence. Some embodiments of the present application can achieve clustering of multiple target objects, with high versatility and high clustering efficiency.

The invention discloses a low-voltage distribution area short-term power load prediction method, system, device and medium, and relates to the field of distribution area load prediction, the technical scheme is characterized in that real-time parameter factors influencing power load characteristics of a low-voltage distribution area are acquired, and the real-time parameter factors comprise a climate factor, an electricity price factor and a time factor; inputting the real-time parameter factor into a pre-constructed load characteristic weight analysis model for analytic hierarchy process to obtain respective first weights of the climate factor, the electricity price factor and the time factor; inputting the first weight into a pre-trained self-organizing mapping neural network, and outputting a clustering center of the power load characteristics; acquiring a daily power load curve of the low-voltage distribution area at the current moment, and clustering the daily power load curve according to the clustering center of the power load characteristics to obtain a clustering result; and inputting the clustering result into a pre-trained BP neural network, and outputting a prediction result of the short-term power load of the low-voltage distribution area.

The present invention relates to the field of image processing technology, and in particular to an intelligent analysis system for obstetric ultrasound examination data, comprising: obtaining an obstetric ultrasound image; obtaining the discreteness of a pixel point according to a preset sliding window of the pixel point, and then obtaining a characteristic value of the pixel point; obtaining the probability that the pixel point is speckle noise according to the difference between the characteristic value of the pixel point and the characteristic values ​​of other pixel points in the neighborhood, and then obtaining an improved metric distance, clustering the obstetric ultrasound image according to the improved metric distance to obtain a number of clustering clusters; obtaining the adjacent clustering clusters of each pixel point, and then obtaining the size of the filter window of each pixel point, and obtaining a denoised image according to the size of the filter window of each pixel point. The present invention can effectively suppress noise and retain the edges between tissue regions by adaptively adjusting the filter window size of each pixel point.

The present invention relates to the technical field of matching enterprise technology needs with intellectual property information, and specifically to an intelligent matching system for enterprise technology needs and intellectual property information. The system includes data storage, natural language processing, machine learning and user interaction modules. The data storage module uses a distributed database to ensure high data availability and scalability; the natural language processing module analyzes and understands the text; the machine learning module implements classification, clustering and matching; and the user interaction module provides a friendly interface. The system improves matching accuracy and efficiency through natural language processing and machine learning technologies, and the distributed database ensures reliable data storage and processing. The system can help enterprises quickly find intellectual property information that meets their needs and promote technological innovation and development.

The invention relates to the technical field of data mining, and discloses a method for improving clustering performance, which comprehensively applies technical means such as subspace learning, nonlinear processing, tensor nuclear norm constraint and consensus similarity learning. Through subspace learning, the method can reveal hidden structures and features in high-dimensional data, and effectively deals with dimension disasters. The nonlinear processing part optimizes nonlinear information in the data and reveals a deep nonlinear structure of the data. By using the tensor nuclear norm constraint, the method not only improves the performance of a clustering algorithm, but also shows a remarkable effect in practical application, for example, the recognition accuracy is improved in face recognition, analysts are helped to more accurately distinguish different consumer groups in the market analysis field and insight into consumer demands, and the market analysis efficiency is improved. Therefore, the accuracy and effectiveness of market analysis are remarkably improved.

The invention provides an abnormal capital collection account group identification, clustering and label labeling method based on capital flow data, and belongs to the technical field of financial risk prevention and control and data mining. A time sequence LSTM-structured self-encoding fusion AI feature extraction module is customized; the method comprises the following steps: capturing time sequence features such as periodic transfer and large-amount concentrated transfer through an LSTM attention layer, extracting structured features such as cross-regional association through an auto-encoder with a risk penalty term, and performing weighted fusion to obtain 12-dimensional AI features; and then semi-supervised K-means is used to identify a suspicious account group, spectral clustering is used to divide a case cluster, a random forest is used to label a'capitator / investor 'label, and the precision is ensured through three-layer verification. The method solves the problems of incomplete artificial feature coverage and poor universal AI adaptability in the prior art, and is suitable for abnormal capital investigation of financial supervision departments.

The application discloses a kind of based on OSM data's expressway interflow node identification and network construction method, first acquisition and pretreatment OSM expressway network and service area data, extract the main line of specified attribute, ramp element and uniform projection coordinate system;Again, by calculating the physical intersection of main line and ramp, the best clustering radius is determined in combination with K distance diagram, interflow area is obtained by DBSCAN clustering and the first type of cutting point is generated, while merging uplink and downlink service area and projecting to generate the second type of cutting point;Based on two kinds of cutting points, the main line section is divided, the end point of the section is matched with the cutting point with an error value of 1.0 meter, interflow, service area and temporary node are generated, and an initial topology graph is constructed;Finally, delete the temporary node connected to only two road sections and simplify the network by merging edges, output structured node and edge data.The application solves the problems of inaccurate interchange identification, road network redundancy and missing service area topology in traditional methods, and realizes the automatic and high-precision construction of expressway logical road network, which is suitable for macroscopic traffic analysis and path planning.

The invention belongs to the technical field of molecular marker technology development and application, and discloses an SSR molecular marker developed based on a mactra veneriformis reference transcriptome sequence, and the SSR molecular marker comprises 16 nucleotide sequences corresponding to microsatellite markers as shown in SEQ ID NO.1-SEQ ID NO.16 in sequence. At present, no report about development of mactra veneriformis microsatellites from genomes and transcriptomes exists, 16 stable microsatellite molecular markers with clear strips and high polymorphism are developed from mactra veneriformis transcriptome data on the basis of a high-throughput transcriptome sequencing technology for the first time, and the microsatellite molecular markers can be used for identifying the mactra veneriformis microsatellites. The method has good practical production and application values in the fields of mactra veneriformis population genetics analysis, genetic relationship identification, high-density genetic map construction and molecular marker-assisted breeding.

The invention provides a daily reactive load curve clustering method and system based on characteristic index dimensionality reduction. The method comprises the following steps: taking a load rate, a highest utilization hour rate, a daily peak-valley difference rate, a peak period load rate, maximum load occurrence time, a flat period load rate, a valley period load rate and minimum load occurrence time as daily reactive load characteristic indexes; carrying out daily reactive load data dimension reduction based on the daily reactive load characteristic indexes, obtaining a daily reactive load characteristic index data dimension reduction matrix, and carrying out weight assignment on the daily reactive load characteristic indexes by adopting an expert weighting method; and carrying out clustering calculation based on a k-means clustering algorithm and outputting a daily reactive load curve clustering result. According to the method, the established daily reactive load characteristic index system is applied to daily reactive load clustering calculation, the calculation amount is effectively reduced, the problem that the clustering effect is not ideal under the high-dimensional condition of data is avoided, curve clustering can be accurately carried out on the daily reactive load, and the method has a guiding effect on making a load scheduling plan and operation control.

The invention relates to the field of data processing, in particular to a multivariate big data optimization storage method, which comprises the following steps of: obtaining a plurality of multivariate data points; obtaining a comparison point, and calculating a correlation description value of each dimension of the comparison point and other multivariate data points; performing clustering processing on all the multivariate data points according to the correlation description values; acquiring a first principal component direction, and acquiring a reference data point according to the first principal component direction; obtaining a unit vector and a modulus length of a description vector of each multivariate data point according to the reference data point; and taking the unit vectors as coding objects, performing Huffman coding on the unit vectors of all the multivariate data points in the category, and storing codes and die lengths corresponding to the unit vectors of the multivariate data points. The data compression amount is improved by selecting proper coding features.

The present invention discloses an environmental monitoring system and method for a modern poultry farm, and relates to the technical field of data processing. The environmental monitoring method for the modern poultry farm comprises: obtaining a temperature change sequence of each breeding area in the poultry farm, the temperature change sequence is used to characterize the temperature change of the breeding area within a preset time period; using each temperature change sequence, determining the clustering coefficient of the regional feature points in each breeding area at the current moment, the regional feature points are characteristic position points preset in the breeding area; clustering the regional feature points according to the clustering coefficient of each regional feature point at the current moment and the distance between each regional feature point, and obtaining at least one current cluster area; using the centroid position change information and regional range change information of the current cluster area, determining the poultry clustering rate of the current cluster area; according to the poultry clustering rate of the current cluster area, determining whether the environmental temperature of the poultry farm is abnormal.

The application discloses a clustering division method and system suitable for magnetic resonance images and a magnetic resonance image information processing device. The method is realized by fusing magnetic resonance image features by using a multi-view algorithm. The method comprises the following steps: acquiring at least two feature similarity matrices according to magnetic resonance image information; fusing all the acquired feature similarity matrices by using a multi-view algorithm to obtain a fusion matrix, wherein the multi-view algorithm is realized by using a similarity network fusion method; and determining a clustering result of the magnetic resonance image information according to the obtained fusion matrix. The clustering division of the magnetic resonance image information is realized by extracting and fusing the magnetic resonance image features, and the clustering result is more accurate and more comprehensive.

The invention discloses a radar point cloud clustering method and system based on vehicle model size adaptive discrimination. The method comprises the following steps: setting parameters of each level of vehicle model size range during clustering, and a corresponding clustering point cloud number threshold value and a clustering speed difference threshold value; obtaining radar point cloud data and sorting to obtain a point cloud data set; traversing reflection points in the point cloud data set in sequence to perform clustering based on the reflection points, and performing point cloud clustering with the vehicle size range expanded step by step by taking the current reflection point as a center to obtain a clustering cluster with the current reflection point as a clustering core; screening out the point cloud data in the clustering cluster from the point cloud data set, and continuing to perform clustering based on the reflection points; and after traversing the point cloud data set, obtaining a plurality of clusters and corresponding vehicle model sizes. According to the method, point cloud clustering with the vehicle type size range expanded step by step is carried out according to the vehicle type size increasing sequence, the point cloud of the large vehicle and the point cloud of the small vehicle can be effectively and accurately clustered, and the problem of splitting of large vehicle clustering is avoided.

A missing multi-view face image clustering method based on double alignment completion belongs to the field of data clustering processing in missing multi-view face image data, and comprises the following steps: firstly, obtaining low-dimensional embedded representation of a missing multi-view face image data sample through an encoding network of an auto-encoder; and obtaining reconstructed sample features through a decoding network, and constraining the missing multi-view face image data and the reconstructed sample features by using a reconstruction loss function. Then, a prototype matrix is constructed by utilizing the potential embedding representation of each view, and similar cluster prototypes between the views are aligned through a prototype comparison loss function; and inputting the potential embedded representation of each view and the corresponding similar cluster prototype into an attention layer, increasing the tightness of similar cluster samples in the views through an attention matrix, and complementing missing views by using similar cluster existing samples among the views. Next, the complemented views are subjected to instance contrast learning to obtain consistency information of multi-view face image data samples among the views, and then common low-dimensional embedding representations of all the views are obtained through a feature fusion layer. And finally, using a small-batch stochastic gradient descent algorithm to optimize the total loss function of the missing multi-view face image clustering method based on double-alignment completion to convergence. And after the model achieves convergence, clustering the common low-dimensional embedded representation by using a k-mean algorithm to obtain a clustering result, thereby achieving the purpose of clustering missing multi-view face image data samples.

The invention discloses a curve clustering-based method and device for analyzing different power utilization modes of industrial loads, a terminal and a medium. The method comprises the following steps of: acquiring power utilization load curve data of each industrial user; preprocessing the load curve data to obtain a complete and available load curve data set; an improved WOA-FCM algorithm is adopted to carry out clustering on the load curve data set; according to a clustering result, distinguishing each industrial power consumption mode; wherein the improved WOA-FCM algorithm comprises the following steps: optimizing the position of each clustering center of a load curve through a WOA whale algorithm, determining the optimal clustering number of the load curve in combination with a set effectiveness index, and optimizing the membership matrix and the clustering center of the current load curve cluster by adopting a fuzzy C-means FCM clustering algorithm, so as to obtain the optimal clustering number of the load curve. And finally obtaining a load curve clustering result of each industrial user. According to the method, more accurate and efficient clustering analysis of different power utilization modes of the industrial load is realized.

The invention relates to the technical field of power system data analysis, in particular to a distributed photovoltaic grid connection line loss diagnosis optimization method, which comprises the following steps of: acquiring power utilization data of a transformer area in real time; analyzing the power consumption data by adopting a line loss detection model constructed by a second-order clustering algorithm to obtain a first line loss diagnosis result; analyzing the first line loss diagnosis result by adopting a line loss abnormity judgment library, and sending out early warning information and a line loss decision when the abnormal data grade meets a preset standard of the line loss abnormity judgment library; according to the line loss decision, optimizing the state of equipment in the transformer area, and analyzing the optimized power consumption data by adopting a line loss detection model to obtain a second line loss diagnosis result; and based on the line loss evaluation index system, comparing and analyzing the first line loss diagnosis result and the second line loss diagnosis result to obtain an optimization effect and generate a line loss management report. According to the method, the problems that the initial classification number is fixed and the effect is poor when a traditional K-means clustering method is used for processing large-scale power data can be solved.

The invention discloses a multi-level multi-attribute decision selection method based on subtraction clustering and combined weighting. The method comprises the following steps: generating a fusion algorithm set; generating a fusion algorithm attribute list and an attribute index matrix; clustering the fusion method by adopting a subtraction clustering algorithm, and outputting to obtain a plurality of subtraction clustering centers; jointly determining the index weight of each type of methods by adopting an analytic hierarchy process and a standard conflict correlation method; and performing multi-level multi-attribute decision selection. According to the method, the uncertainty in the decision making process is reduced, the weight of the method can be updated online when a new method is added or corrected, so that the method set and the whole framework are updated, the effect of decision making selection on the data set can be achieved only by calculating relevant parameters of the to-be-decided algorithm data set in the actual application process, and the decision making efficiency is improved. And the method has relatively high robustness and scene adaptability.

The present application relates to the field of data processing technology, and provides a method, device, electronic device and readable storage medium for analyzing public opinion data. The method includes: obtaining public opinion data, and performing streaming clustering on the public opinion data to obtain the cluster to which the public opinion data belongs; obtaining the attention index value of the cluster based on the attribute feature data of each target text of the cluster; performing public opinion analysis on the public opinion events corresponding to the cluster based on the large language model to obtain the public opinion summary corresponding to the cluster; judging and generating the event category corresponding to the cluster based on the large language model; when the attention index value of the cluster is greater than or equal to the attention index value threshold corresponding to the event category corresponding to the cluster, the event category corresponding to the cluster, the attention index value of the cluster and the public opinion summary corresponding to the cluster are sent to the terminal device. The method solves the problems of low generalization and inaccurate analysis of public opinion data in the prior art, and improves the efficiency and accuracy of public opinion monitoring and management.

This application discloses a method for determining developmental trajectories based on single-cell multi-omics clustering, belonging to the field of biomedical data mining technology. The method includes: acquiring multi-omics data of single cells from the same tissue; determining the potential representation of each single cell in different omics based on feature encoding technology; constructing a K-nearest neighbor graph for each omics based on the distance between single cells; and determining the corresponding multi-order similarity matrix; using the multi-order similarity matrix to complete the missing potential representations of single cells, obtaining the complete potential representation of each omics; performing cluster analysis on each omics to obtain single-cell clustering results; weighted fusion of the potential representations of each single cell in different omics to obtain a comprehensive potential representation; and analyzing the developmental trajectory of single cells based on the comprehensive potential representation. This application can stably and accurately cluster single cells under conditions of missing single-cell omics or significant differences in omics quality, thereby accurately determining the developmental trajectory of single cells.