70 results about "Fuzzy clusters" patented technology

The invention discloses a territorial space layout adjustment decision method based on planning conflict identification, and belongs to the technical field of territorial space planning and management. Comprising the following steps: constructing a space-time conflict feature vector library containing multi-dimensional attributes; generating a semantic knowledge graph containing a causal relationship, a sequential relationship, a spatial relationship and an affiliation relationship; accurate prediction of influence intensity and duration of the conflict state along with time lapse is realized; trend analysis and risk level evaluation of a cross-scale conflict propagation path are realized; automatically identifying dominant conflicts in the current layout and potential conflicts based on propagation prediction, and intelligently classifying composite conflicts by using fuzzy clustering; a staged differentiation conflict intervention strategy recommendation scheme is automatically generated; solving a territorial space layout adjustment scheme; judging whether the layout adjustment scheme meets a preset conflict relieving standard or not;

The invention provides an image noise mark feature selection method and system, a storage medium and a computer. The method comprises the steps of obtaining a to-be-processed image noise mark data set; embedding a sample set in the image noise mark data set into a multi-granularity fuzzy cluster to construct a dynamic fuzzy membership evaluation matrix; dynamically evolving a multi-level high-precision granular ball cluster; obtaining mark distribution with high identification degree; constructing a rough perception feature evaluation framework based on granular ball topology driving, extracting decision equivalence classes by combining rough set upper and lower approximation and extended positive domain models, and determining and measuring the contribution degree of each feature to a decision system by fusing multi-granular-ball decision boundary information based on a dependency degree quantitative model; a particle and ball structure consistency verification mechanism is introduced, and multi-level evaluation is carried out on the importance of the features through dependency and consistency. According to the method, the optimal feature subset with strong anti-noise performance and high discrimination capability is obtained, and stable and efficient input support is provided for a subsequent image noise mark learning model.

The invention relates to the technical field of lithium battery energy storage and temperature control, in particular to an electronic radiator intelligent temperature control system based on a lithium battery energy storage cabinet. The system comprises an edge temperature sensing module, a cloud thermal management strategy module, a cloud clustering analysis module, a thermal field topological optimization module, a multi-source thermal feature fusion module, a dynamic temperature control strategy generation module and a grading strategy issuing execution module. An edge temperature sensing module collects real-time temperature data of multiple monitoring points and an edge side preliminary temperature control scene judgment result; the cloud heat management strategy module receives battery pack topology connection data; the cloud clustering analysis module carries out fuzzy clustering on the data to obtain a temperature state and scene discrimination clustering center; the thermal field topological optimization module optimizes a temperature state clustering center in combination with topological data; the multi-source thermal feature fusion module fuses the two clustering centers to form multi-source fusion feature representation; the dynamic temperature control strategy generation module determines a temperature control strategy identifier accordingly, and the grading strategy issuing execution module matches and distributes a gradient temperature control strategy to a heat dissipation actuator.

The invention relates to the technical field of aero-engine fault prediction, in particular to an aero-engine residual life prediction method based on fuzzy modal regression SCN, and the method comprises the steps: obtaining multi-source monitoring data of an aero-engine under different working conditions; performing fuzzy clustering on the training sample set based on a fuzzy C-means algorithm; training the SCN structure according to the subspace training sample; optimizing the local prediction sub-model based on a modal regression model; according to a generalized maximum correlation entropy loss function, performing robustness optimization on the local prediction sub-model after modal regression in combination with a semi-quadratic optimization strategy and a sparse regular term; predicting a to-be-tested aero-engine sample according to the local aero-engine residual life prediction model; and fuzzy weighted fusion is carried out on the local residual life prediction value of the aero-engine. According to the method, the prediction information of each modal subspace can be integrated, the multi-modal degradation behavior is comprehensively represented, and the accuracy of prediction of the residual life of the aero-engine is remarkably improved.

The invention provides an industrial air conditioner and waste heat recovery collaborative intelligent optimization control method and related equipment, and is applied to the technical field of data processing. Industrial air conditioner and waste heat recovery system data are fused through the Internet of Things and edge calculation, a standardized parameter sequence is generated through wavelet noise reduction and is converted into an energy flow network atlas, high-energy-consumption working conditions are subdivided through fuzzy clustering, and a dynamic energy efficiency coupling model is constructed. And based on the model, constructing an optimization equation by using an improved particle swarm algorithm, and obtaining the energy consumption of the key node in combination with an energy gradient utilization matrix. And an energy-saving potential coefficient is generated through the multi-dimensional feature matrix, a production plan grouping operation mode is combined, a self-adaptive optimization control model is constructed by using deep reinforcement learning, and finally a real-time regulation and control instruction and an energy efficiency scheme are generated to realize system collaborative optimization.

The invention provides a driving anti-slip optimization control method based on vehicle-road cooperation road fusion estimation, and the method comprises the following steps: S1, fusing vehicle-mounted and roadside visual recognition information, and obtaining a vehicle-road cooperation road state recognition result through employing a weighted voting method in combination with road invariant hypothesis and historical information assistance judgment; s2, adaptively adjusting parameters of a dynamic estimator by using a pavement state recognition result, establishing a sample set based on a pavement fusion estimation result, introducing a fuzzy clustering algorithm of particle swarm optimization to realize self-evolution optimization of the structure and the parameters of the estimator, and outputting a fused attachment coefficient estimation result; and S3, designing a slip rate and wheel acceleration combined control strategy to deal with potential slip in advance, and outputting a control torque result of a slip rate controller. Compared with the prior art, the method integrates visual perception and dynamic information to realize accurate estimation of the pavement state, supports advanced intervention of the slip rate controller before slip, and improves the driving stability of the whole vehicle.

The invention discloses an electric vehicle charging load prediction method based on an optimized fuzzy neural network, and the method comprises the steps: obtaining a plurality of historical load data of an electric vehicle, and processing all historical load data to generate a multi-dimensional feature data set; dividing an input feature space for the multi-dimensional feature data set based on an improved fuzzy clustering algorithm to obtain a clustering result and an activation intensity-contribution degree two-dimensional evaluation index; iteratively optimizing FNN parameters based on an improved adaptive differential evolution algorithm; constructing a rule contribution degree analysis model based on an orthogonal experimental design, setting a quantitative truncation threshold with the cumulative interpretation degree greater than or equal to 85%, and realizing self-adaptive compression of the scale of the rule base; an activation intensity threshold value dynamic calculation method is used, low-efficiency rules are automatically identified through sliding window statistics, and rule pruning is achieved; and outputting a load prediction result, and applying non-negativity and peak constraint to a prediction value.

The invention discloses an elevator maintenance quality evaluation method based on an SVR-NSGA-I algorithm. The method comprises a data acquisition stage, a data preprocessing stage, an SVR model training stage, an NSGA-I optimization stage and an evaluation system construction stage. Compared with the prior art, the method has the advantages that PSO optimized SVR and adaptive NSGA-I are deeply fused, the multi-target optimization problem of elevator maintenance is solved, an isolated forest-KNN combined preprocessing method is provided, high-dimensional, nonlinear and noisy maintenance data are effectively processed, a time decay factor and fuzzy clustering are introduced, and the maintenance efficiency is improved. The real-time dynamic adjustment of the evaluation standard is realized, and the model transparency and practicability are improved in combination with SHAP value analysis and visual decision support.

The invention provides a rotational symmetry point cloud registration method and device based on fuzzy clustering and a storage medium, and the method comprises the steps: extracting inner points through employing an algorithm combining the minimum median square and random sampling consistency, and estimating a rotating main axis through principal component analysis; realizing point cloud global coarse registration based on a rotating main shaft; fuzzy clustering is carried out on the fixed point cloud to construct local geometric description, a progressive constraint mechanism and a directional search strategy based on a main shaft are introduced, and fine registration is completed in combination with branch and bound and gradient descent algorithms. According to the embodiment of the invention, the registration precision and efficiency of point clouds with rotational symmetry, such as a computer-aided design (CAD) model, can be effectively improved.

The invention provides a cartridge case precision forming process abnormal working condition monitoring and early warning method. Diffusion distance fuzzy clustering (DDFCM) and dynamic internal principal component analysis (DiPCA) technologies are fused. The method comprises the following steps: firstly, performing fuzzy clustering on process data based on a diffusion distance by adopting a DDFCM method, automatically identifying key sample groups under different working conditions, and constructing a representative local training data set; thirdly, for each clustering sample group, constructing a dynamic monitoring model by using a DiPCA technology, effectively extracting potential variables containing time lag features, and capturing a dynamic change mode of a process; in a new data acquisition stage, firstly, the membership degree of a new sample in each cluster is calculated through DDFCM, and the new sample is distributed to a corresponding local DiPCA model for state discrimination according to a maximum membership degree principle. According to the method, the time sequence dynamic change in the process can be fully captured, the local characteristics in different production states are also considered, and a powerful guarantee is provided for the production stability and the product quality of the cartridge case coating, rolling and cutting process.

A computing entity obtains product information corresponding to a product and defines a multi-region metric space vector based on the product information. The multi-region metric space vector is a vector within a multi-region metric space that comprises a first region corresponding to supply chain / component information, a second region corresponding to assembly / fabrication information, and a third region corresponding to usage / usage environment information. Each of the first, second, and third regions are multi-dimensional. The computing entity processes the multi-region metric space vector using a component, assembly, and usage perspicacity model configured to define fuzzy clusters within the multi-region metric space; determines a parameter corresponding to the product based on at least one fuzzy cluster to which the perspicacity model assigned the multi-region space vector; and provides or causes providing of (a) a visual / audible representation of the parameter or (b) a machine-readable representation of the parameter as input to an application / program.

The present invention relates to the technical field of economic resource management, and discloses an economic resource management optimization method based on intelligent decision-making. The method first collects multi-source heterogeneous data of the economic system, including resource stocks, demand fluctuations, etc. Then, based on multi-dimensional feature analysis, the dynamic resource pool is divided, and a nonlinear optimization model is constructed to predict changes in resource supply and demand to generate an allocation plan. Then, a multi-stage decision tree is used to optimize the allocation path, and the model parameters are updated through an adaptive learning mechanism according to market feedback and changes in constraints, and the allocation plan is corrected in real time. In addition, key technical details such as the elastic quota adjustment formula and the fuzzy clustering algorithm membership function are also given. The present invention can effectively integrate complex data, scientifically dispatch resources, accurately predict supply and demand, optimize allocation paths, adapt to environmental changes, significantly improve the efficiency and benefits of economic resource management, and provide scientific and reasonable resource management decision support for economic entities.

This invention discloses a robust fusion method for large models based on semantically aligned fuzzy clustering. The method includes: first, obtaining a sequence of probability distribution vectors from the outputs of multiple heterogeneous large-scale pre-trained models for the samples to be processed; then, introducing non-negative reliability weights to weight this sequence to obtain an aggregated probability matrix, and constructing a graph Laplacian matrix accordingly; second, approximating the aggregated probability matrix into a fuzzy membership matrix and a semantic prototype matrix, constructing a cost objective function by combining the graph Laplacian matrix, and iteratively updating each matrix and weight until convergence; finally, solving for the maximum value of the converged fuzzy membership matrix to obtain the sample prediction category result. This invention can effectively suppress the influence of inferior models in unsupervised environments, significantly improving the accuracy and robustness of fusion prediction.

This application discloses a method for determining the threshold of watershed ecological environment quality based on the PSR model, belonging to the field of watershed ecological assessment technology; it includes: determining the evaluation index items of the target watershed; arranging and combining all possible score values ​​of each index to obtain all score combinations; calculating the comprehensive score of each combination to generate a full sample space; performing fuzzy clustering on the full sample space, iteratively determining the membership degree and cluster center value of each sample to multiple preset cluster categories; sorting the cluster center values ​​of each cluster category in ascending order to form multiple health levels; classifying each sample to the cluster category with the largest membership degree based on the principle of maximum membership degree to obtain the sample set corresponding to each health level; fitting continuous distribution density curves to the sample sets of each health level, and determining the intersection of the distribution density curves of adjacent health levels as the critical threshold between adjacent health levels; and determining the health level of the target watershed based on each critical threshold.

This invention proposes a mobile swarm intelligence sensing task allocation method and system based on fuzzy clustering and dynamic scheduling. First, fuzzy clustering is used to mine the latent spatial structure of tasks, and tasks are classified into fuzzy clusters with membership attributes through soft partitioning. After obtaining the spatiotemporal distribution characteristics of tasks, a benefit matrix including spatiotemporal feasibility and time cost is constructed, and the Hungarian algorithm is used to achieve a globally optimal match between worker and task clusters. For the online execution phase, a dynamic scheduling mechanism including a primary task pool and a collaborative task pool is designed. Workers select tasks based on a hybrid priority index combining time urgency, spatial proximity, and cluster membership. This invention solves the problem of balancing global efficiency and local flexibility in large-scale task allocation through a hierarchical mechanism of "macro-partitioning-meso-matching-micro-scheduling," significantly improving the overall task completion rate while effectively reducing the system's time resource cost.

The invention discloses a small sample hyperspectral image classification method and device based on fuzzy contrast graph convolution. The method comprises the following steps: determining membership information, fuzzy node initial feature vectors and similarity of each initial graph node corresponding to each fuzzy cluster based on a hyperspectral image; generating a fusion weight, a fusion feature and a target feature according to convolution features output by convolution of different layers of fuzzy images; performing linear transformation operation and probability distribution operation on the target features in sequence to obtain the probability that the corresponding initialized graph node belongs to each category; and performing argmax operation on the probability to obtain a category index corresponding to the hyperspectral image. The uncertainty of pixels is effectively processed through fuzzy learning, interference of neighborhood pixels is weakened, the uncertainty between the pixels is better processed, and meanwhile richer and more expressive feature representation is obtained. The method has higher classification effect and robustness, and plays an important role in multiple fields of mineral exploration, environment monitoring, forest management, precision agriculture and the like.

The application discloses a kind of based on fuzzy clustering and dynamic weighting fusion's arrester operating state data optimization method, this method includes: step 1: the extraction and normalization processing of arrester full current signal and frequency domain feature;Step 2: setting the density peak method initialization cluster center of integrated decision value condition;Step 3: using Gaussian kernel function to the data after processing Nonlinear mapping, construct the improved fuzzy clustering (IFCM) objective function with regular term and solve;Step 4: combined with fuzzy membership and dynamic weighting fusion mechanism of Mahalanobis distance constructs data comprehensive score model;Step 5: based on adaptive threshold screening abnormal data.The method effectively overcomes the sensitivity of traditional fuzzy clustering algorithm to initial value and the limitation of Euclidean distance, can automatically identify and eliminate strong volatility, low-quality data points disturbed, significantly improve the accuracy and robustness of health assessment model.

The invention discloses a double-view 3D medical image segmentation method based on fuzzy clustering Mama driving, and the method comprises the steps: employing a segmentation network 1 (2) and evaluation network 2 (1) cross learning mechanism based on a Mama-DF3D model, training and improving the model performance, carrying out the 3D medical image segmentation, and receiving 3D medical image data; inputting the equal-proportion annotated data and the unannotated data into a Mamba-DF3D model respectively to carry out network segmentation and network evaluation parameter training optimization; and carrying out 3D medical image segmentation by using the trained and optimized Mmba-DF3D model. According to the model, the dependence on a large amount of labeled data is effectively reduced, the cost of obtaining high-quality labels is remarkably reduced, the working efficiency is improved, medical resources are more reasonably distributed, the popularization of a new technology is promoted, a new thought and a framework are provided for subsequent research, the innovation of related fields is promoted, and wider economic benefits are brought.

The invention provides a mobile crowd sensing task allocation method and system based on fuzzy clustering and dynamic scheduling. The method comprises the following steps: firstly, mining a potential space structure of a task by using fuzzy clustering, and classifying the task into fuzzy clustering with membership attribute through soft division; after spatio-temporal distribution characteristics of tasks are obtained, a benefit matrix containing spatio-temporal feasibility and time cost is constructed, and global optimal matching between workers and task clusters is realized by adopting a Hungary algorithm. For an online execution stage, a dynamic scheduling mechanism comprising a primary task pool and a cooperative task pool is designed, and a worker screens tasks according to a mixed priority index combining time urgency, spatial proximity and clustering membership. Through a hierarchical mechanism of'macroscopic division-mesoscopic matching-microscopic scheduling ', the problem that global efficiency and local flexibility in large-scale task allocation are difficult to consider at the same time is solved, and the time resource cost of the system is effectively reduced while the overall task completion rate is remarkably improved.

This invention relates to an adaptive grouping method for multi-role students based on reinforcement learning and linear programming, applicable to the field of teaching and education. It employs fuzzy clustering of students based on their multidimensional features, followed by reinforcement learning to filter the grouping of large student populations, reducing the problem size. Linear programming is then used to rate the filtered rules, which are fed back into a deep network for further learning. Finally, an effective grouping scheme is output. The advantages of this invention are: considering that students may have multiple roles, this method performs fuzzy clustering based on their multidimensional features, then uses reinforcement learning to filter the grouping of large student populations, significantly reducing the problem size. Linear programming is then used to rate the filtered rules, which are fed back into a deep network for further learning, ultimately outputting an effective grouping scheme for the current student population. This approach can achieve good grouping results in a relatively short time.

The application relates to a self-step fuzzy clustering method and system for incomplete multi-view medical images, and relates to the field of intelligent analysis of medical images. The application solves the problems that existing multi-view clustering technologies cannot effectively process missing partial view data, fuzzy organization structure and lack a progressive dynamic sample selection mechanism. The method comprises the following steps: collecting multi-view medical image data, constructing a view missing label matrix, and performing spatial registration and feature standardization processing; based on a K nearest neighbor neighborhood weighting reconstruction strategy of a feature space, view features with missing data are completed; a standard multi-view fuzzy clustering is adopted, a joint fuzzy clustering objective function fusing a view missing label is constructed, an adaptive view weight distribution strategy and a self-step learning strategy are adopted, and the confidence of a clustering model of fuzzy membership entropy is defined; an alternating optimization strategy is used to update the membership, the clustering center and the view weight of the clustering model, and the self-step fuzzy clustering for incomplete multi-view medical images is completed.

The invention discloses a digital marketing recommendation method and system based on big data, and relates to the technical field of commodity recommendation. The method comprises the steps of obtaining original data of an e-commerce platform, obtaining a structured data set through preprocessing, performing fuzzy clustering on a scoring matrix, generating and iteratively updating a membership matrix, reducing the influence of initial clustering deviation, calculating clustering attribution in combination with user geographic information, and correcting scores according to the clustering attribution to obtain a final scoring matrix. Inputting the matrix and the user and commodity feature vectors into a recommendation model, and outputting a commodity recommendation result; clustering multiple membership degrees of the users by adopting fuzzy clustering, and adapting to complex overlapping characteristics of user preferences under sparse data of the distributors; the initial deviation is reduced by iteratively optimizing the membership degree, attribution is determined in combination with geographic information, and the clustering fitness and distinction degree are improved; the commodity recommendation model effectively solves the problem of low accuracy of the commodity recommendation result under the condition of reducing the calculation complexity.

The invention relates to a power distribution network fault monitoring method and monitoring equipment, and the method comprises the following steps: data collection: collecting data information in a power distribution network line, and achieving the automatic fault recording; receiving data: receiving various data and fault recording information, and uploading the data for analysis and processing; data preprocessing: carrying out standardization processing on related data by adopting a similar maximum standardization method; current data clustering calculation is carried out; a C mean value fuzzy clustering algorithm is adopted to realize power distribution network line load prediction; current-day load mode matching: determining a current-day load affiliation category by adopting a maximum membership degree method; mismatch degree monitoring: defining a mismatch degree index considering membership degree and Euclidean distance factors as a fault evaluation criterion; fault type research and judgment: determining the fault type of the line which is judged to be faulty; fault positioning: accurately positioning a fault position by adopting a single-end distance measurement algorithm; the method has the advantages of fault monitoring analysis based on big data, fault evaluation criterion definition and accurate positioning.

The invention provides a voltage-withstanding flashover ultrasonic signal determination method and a related device. The method comprises the following steps: acquiring a discharge data set; first feature information is determined based on the discharge data set and a first model, second feature information is determined based on the discharge data set and a second model, the first model is an unsupervised dimensionality reduction feature extraction model and is used for analyzing principal components of the discharge data set, and the second model is a supervised dimensionality reduction feature extraction model and is used for analyzing principal components of the discharge data set. The feature classification module is used for performing feature classification; determining a double-branch prediction probability based on the first feature information, the second feature information, and an accurate clustering sub-model and a fuzzy clustering sub-model of a pre-constructed clustering model; and determining the withstand voltage flashover ultrasonic signal type of the target network based on the double-branch prediction probability. In this way, high-precision and high-robustness recognition of multi-type flashover and noise in the GIS is achieved, the feature utilization rate can be increased, the model generalization ability can be improved, and the calculation efficiency can be improved.

This invention provides a soft measurement method and system for fermentation processes based on continuous membership degrees. The method includes acquiring and preprocessing historical fermentation process data to construct a training dataset; performing fuzzy clustering on the historical process variable data to identify K local evolution stages of the fermentation process and calculating initial membership degrees; performing time-series continuous processing on the initial membership degrees to generate a historical continuous membership degree sequence; constructing K sub-models, training each sub-model, and saving cluster center parameters and time-series continuous calculation parameters; synchronously inputting the real-time process variable sequence into each sub-model to obtain local soft measurement prediction values; calculating the current continuous membership degree value as a dynamic weight, and weighting and fusing the local soft measurement prediction values ​​to obtain a global soft measurement prediction result. The beneficial effect of this invention is that by replacing the traditional hard switching flag with continuous membership degrees, the weights of multiple models change continuously and gradually over time, eliminating prediction steps and high-frequency oscillations in the stage transition zone.

An embodiment of the present invention provides a method and device for generating bandwidth data prediction results, which obtains historical bandwidth usage data and generates a timestamp corresponding to the historical bandwidth usage data; generates multiple historical data time series for the historical bandwidth usage data based on the timestamp; divides the historical bandwidth usage data into multiple bandwidth fuzzy clusters through the multiple historical data time series; establishes a prediction data cluster attribution model based on the multiple bandwidth fuzzy clusters; determines neural network initialization parameters, and generates a bandwidth prediction model through the prediction data cluster attribution model and the neural network initialization parameters; and generates a bandwidth prediction result based on the bandwidth prediction model, thereby improving the efficiency and accuracy of bandwidth data prediction, reducing the maintenance cost of underlying equipment, and further improving the normal brightness rate of pass codes.

Embodiments of the present application disclose a person care method and device, computer equipment and a storage medium. The method comprises: obtaining log data of each door magnetic sensor arranged at a residence of a monitored person to obtain original sequence group data; performing adaptive fuzzy clustering on the original sequence group data to obtain a plurality of sequence group sets; determining a clustering center of each sequence group set and a maximum distance of the clustering center, and setting the maximum distance of the clustering center as a time deviation threshold; obtaining a monitoring signal collected by each door magnetic sensor in real time; determining whether a distance between current monitoring data and any sequence group set is within the corresponding time deviation threshold; if yes, determining that the current monitoring data are matched with the plurality of sequence group sets successfully; determining a behavior category of the monitored person according to the matched sequence group set; and if not, generating a remote reminding information. The method of the embodiments of the present application can improve the safety care efficiency of the monitored person.

The application provides a multi-modal data security analysis technology based on a deep fuzzy wavelet learning model. The deep fuzzy wavelet learning model is widely used in the field of security analysis due to its self-adaptive ability in complex data environment. The method improves the deep fuzzy wavelet learning framework and applies it to multi-modal data security analysis. The model first uses the method combining wavelet transform and fuzzy clustering to perform feature decomposition and noise reduction processing on multi-modal data (text, speech, video, image), extracts key semantic, timing and spatial features, and provides high-dimensional information representation for subsequent security analysis. Secondly, a deep fuzzy wavelet neural network is constructed, the extracted features are input into the adaptive fuzzy decision layer, and the wavelet learning model is combined to capture the spatio-temporal correlation features, improve the perception ability of the model to complex attack patterns. Finally, the security analysis result is strengthened by using the space search optimization algorithm, so as to effectively detect malicious samples, identify backdoor attacks, and improve the defense ability against attacks, and finally realize accurate multi-modal data security evaluation.

The invention provides a load prediction method and system based on virtual power plant operation, and the method comprises the steps: building a hybrid prediction model based on an ant colony optimization algorithm dynamic weighted combination LSTM time sequence sub-model, an XGBoost gradient promotion sub-model and a fuzzy clustering sub-model; a main deviation source is positioned through an error propagation network, and hybrid prediction model reconstruction is triggered according to the real-time operation state of the virtual power plant; calculating a load predictability index of the reconstructed hybrid prediction model; and correcting a load prediction result of the hybrid prediction model according to the load predictability index to obtain a final load prediction value. According to the method, the sub-models are dynamically weighted and combined through the ant colony optimization algorithm to construct the hybrid prediction model, a deviation source is positioned through an error propagation network, the model is reconstructed according to the real-time operation state of the power plant, the weights of the sub-models are adjusted, and the predictability index of the model is calculated to evaluate the reliability of the model so as to correct the result. And the accuracy of virtual power plant load prediction is greatly improved.

The invention discloses a lightning arrester operation state data optimization method based on fuzzy clustering and dynamic weighted fusion. The method comprises the following steps: step 1, extracting and normalizing a total current signal and frequency domain characteristics of a lightning arrester; step 2, initializing a clustering center by a density peak method for setting a comprehensive decision value condition; 3, performing nonlinear mapping on the processed data by adopting a Gaussian kernel function, constructing an improved fuzzy clustering (IFCM) objective function with a regular term, and solving the objective function; 4, constructing a data comprehensive scoring model in combination with a dynamic weighted fusion mechanism of fuzzy membership and mahalanobis distance; and 5, screening abnormal data based on a self-adaptive threshold value. According to the method, the defects that a traditional fuzzy clustering algorithm is sensitive to an initial value and the Euclidean distance is limited are effectively overcome, low-quality data points which are high in volatility and interfered can be automatically recognized and eliminated, and the accuracy and robustness of a health assessment model are remarkably improved.