Hierarchical measurement time sequence analysis method and device, electronic equipment and storage medium

Abstract

The invention provides a hierarchical measurement time sequence analysis method and device, electronic equipment and a storage medium, and relates to the technical field of data analysis, and the technical scheme is characterized in that the method comprises the following steps: obtaining total node time sequence data and child node time sequence data of a system; mapping the total node time sequence data and the child node time sequence data to the same space to obtain consistency representation; and performing analysis according to the consistency representation to obtain an analysis result. The hierarchical measurement time sequence analysis method and device, the electronic equipment and the storage medium provided by the invention have the advantages that multi-level data information can be fully utilized, and the accuracy and effect of clustering analysis are improved.

Description

Technical Field

[0001] The present application relates to the technical field of data analysis, and in particular to a hierarchical quantitative time series analysis method, device, electronic device and storage medium. Background Art

[0002] Time series clustering is a data analysis technology that mainly classifies data by mining the sequence similarities in time series data sets. By mining meaningful features in the sequence and performing cluster analysis, the power consumption pattern of electrical equipment can be clearly revealed, thus providing accurate data support for energy management and optimization.

[0003] Currently, most time series clustering methods extract features from the time series data of the target component, including the use of dimensionality reduction techniques and spectral analysis, and then perform clustering analysis from a single perspective based on these features and the generated pseudo-labels.

[0004] Most existing time series clustering algorithms are based on a single perspective for analysis, which often results in data details being ignored. For example, for a device composed of multiple sub-nodes, the overall power consumption cannot accurately reflect the specific power consumption behavior of each sub-node, and there may be significant power consumption differences between different sub-node components. Existing single-perspective algorithms ignore the correlation between the various levels of hierarchical metering components, so they are insufficient in identifying the power consumption patterns of each component, thus limiting the improvement of clustering performance.

[0005] In view of the above problems, this application proposes a new solution. Summary of the invention

[0006] The purpose of the present application is to provide a hierarchical quantitative time series analysis method, device, electronic device and storage medium, which have the advantages of being able to fully utilize multi-level data information and improve the accuracy and effect of cluster analysis.

[0007] In a first aspect, the present application provides a hierarchical measurement time series analysis method for performing time series analysis on a system composed of multiple sub-nodes. The technical solution is as follows: include: Obtaining total node time series data and sub-node time series data of the system; Mapping the total node time series data and the sub-node time series data to the same space to obtain a consistent representation; An analysis result is obtained by performing analysis according to the consistency expression.

[0008] Furthermore, in the present application, the steps of the method also include: The projection matrix in the process of mapping the total node time series data and the sub-node time series data into the same space to obtain a consistent representation is constrained by orthogonality and standardization constraints.

[0009] Further, in the present application, the step of performing analysis according to the consistency representation to obtain an analysis result includes: According to the consistency representation, the total node time series data and the sub-node time series data, acquiring a relationship structure between the total node time series data and the sub-node time series data; An analysis result is obtained by performing analysis according to the consistency representation and the relationship structure.

[0010] Further, in the present application, the step of acquiring the relationship structure of the total node time series data and the child node time series data according to the consistency representation, the total node time series data and the child node time series data includes: Constructing a graph structure according to the total node time series data and the sub-node time series data to obtain the relationship structure; Also includes: Performing regularization processing on the graph structure to obtain a regularization term of the graph structure; Performing regularization processing on the consistency representation to obtain a regularization term of the consistency representation; The relational structure is obtained according to the graph structure, the regularization term of the graph structure, and the regularization term of the consistency representation.

[0011] Further, in the present application, the formula for mapping the total node time series data and the sub-node time series data to the same space to obtain a consistent representation is: ; in, Indicates The data matrix input from each perspective includes the total node perspective and the sub-node perspective, corresponding to the total node time series data and the sub-node time series data; Indicates the summary point perspective; Indicates the child node perspective; Indicates The projection matrix of the view input; Indicates consistency representation; represents the regularization parameter; is the square of the Frobenius norm; The orthogonality and normalization constraints are: ; in, and Both represent the identity matrix; represents the dimension of the identity matrix; Represents a transpose operation.

[0012] Further, in the present application, the formula for obtaining the relational structure according to the graph structure, the regularization term of the graph structure and the regularization term of the consistency representation is: ; in, and It represents different time segments; It indicates the total number of time segments; It represents the similarity matrix; It means for The regularization parameter of ; represents the transpose operation, that is, It means that The transpose of It means for The regularization parameter of ; It represents the trace operation; is the square of the Euclidean norm; It also includes constraints, which are: .

[0013] Further, in the present application, the step of performing analysis according to the consistency representation to obtain an analysis result includes: constructing a mathematical model based on the consistency representation; Performing analysis according to the mathematical model to obtain analysis results; The mathematical model is: ; , ; The step of analyzing according to the mathematical model to obtain the analysis result comprises: from , , In the process of changing one parameter in turn and fixing other parameters for alternate optimization, it includes: fixed , ,renew ; fixed , ,renew ; fixed , ,renew .

[0014] In a second aspect, the present application also proposes a hierarchical measurement time series analysis device for performing time series analysis on a system composed of multiple sub-nodes, including: An acquisition module, used to acquire the total node time series data and the sub-node time series data of the system; A mapping module, used for mapping the total node time series data and the sub-node time series data to the same space to obtain a consistent representation; An analysis module is used to perform analysis according to the consistency representation to obtain an analysis result.

[0015] In a third aspect, the present application further proposes an electronic device, comprising a processor and a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps in the above method are executed.

[0016] In a fourth aspect, the present application further proposes a storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps in the above method are executed.

[0017] From the above, it can be seen that the hierarchical metrological time series analysis method, device, electronic device and storage medium provided by the present application fully utilize multi-level data information by mapping the time series data of the total node and the sub-node to the same space and performing analysis based on consistent representation, thereby solving the problem of single-perspective analysis in the prior art that data details are ignored, and has the advantage of being able to fully utilize multi-level data information and improve the accuracy and effect of cluster analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A flowchart of a hierarchical quantitative time series analysis method provided for this application.

[0019] Figure 2 A schematic diagram of the structure of a hierarchical measurement time series analysis device provided in this application.

[0020] Figure 3 A schematic diagram of the structure of an electronic device provided in this application.

[0021] In the figure: 210, acquisition module; 220, mapping module; 230, analysis module; 310, processor; 320, memory. DETAILED DESCRIPTION

[0022] The technical solutions in the present application will be clearly and completely described below in conjunction with the drawings in the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the present application for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0023] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0024] Please refer to Figure 1 In the first aspect, the present application provides a hierarchical measurement time series analysis method for performing time series analysis on a system composed of multiple sub-nodes. The technical solution includes: S110, obtaining the total node time series data and the sub-node time series data of the system; S120, mapping the total node time series data and the sub-node time series data to the same space to obtain a consistent representation; S130, performing analysis according to the consistency representation to obtain analysis results.

[0025] Among them, the total node time series data refers to the time series data of the entire system, which can be achieved by using indicators such as the system's overall energy consumption and overall output.

[0026] The sub-node time series data refers to the time series data of each sub-component or sub-system in the system, which can be specifically realized by indicators such as energy consumption and output of each sub-system.

[0027] Among them, consistent representation refers to the unified representation obtained by mapping time series data of different levels to the same space, which can be achieved by using dimensionality reduction technology or feature extraction methods. In some specific implementations, the mapping referred to here is a low-dimensional projection, that is, the total node time series data and the sub-node time series data are projected into the same low-dimensional space to obtain a consistent representation.

[0028] The core innovation of this application is to propose a multi-level, multi-perspective time series analysis method. This method overcomes the limitations of traditional single-perspective analysis by simultaneously considering the system's total node and sub-node data and mapping them to the same space for analysis. This method can more comprehensively capture the relationship between each level of the system and improve the accuracy and comprehensiveness of time series analysis.

[0029] How this application works can be divided into three main steps: data acquisition, data mapping, and data analysis.

[0030] In the data acquisition step, first collect the time series data of the master node and all sub-nodes. These data may come from different sensors or data sources, including energy consumption data, production data, etc. The data collection frequency and time span need to be set according to the specific application scenario to ensure the representativeness and integrity of the data.

[0031] The data mapping step is the key innovation of this application. In this step, the time series data of the total node and the child node are mapped to the same feature space. This process can be achieved through a variety of mathematical methods, such as principal component analysis (PCA), autoencoders, etc. The purpose of mapping is to reduce the data dimension while retaining the key information and structure in the data. In the mapping process, the weights of data at different levels may need to be considered to balance the importance of overall and local information.

[0032] The data analysis step is based on the mapped consistent representation. This step can use a variety of analysis methods, such as cluster analysis, anomaly detection, pattern recognition, etc. Since the consistent representation contains multi-level information, the analysis results can reflect the relationship and influence between the various levels of the system.

[0033] The reason for choosing this multi-level analysis method is that it can consider the overall performance and local characteristics of the system at the same time, avoiding the information loss that may be caused by single-perspective analysis. By mapping data at different levels into the same space, this method not only reduces the complexity of the data, but also retains the relationship information between levels, thereby improving the accuracy and comprehensiveness of the analysis.

[0034] Furthermore, in some embodiments, the method further comprises: The projection matrix in the process of mapping the total node time series data and the sub-node time series data into the same space to obtain a consistent representation is constrained by orthogonality and standardization constraints.

[0035] The technical solution proposed in this application can effectively improve the accuracy and effectiveness of mapping the total node and sub-node time series data to the same space to obtain a consistent representation by imposing orthogonality and standardization constraints on the projection matrix. Orthogonality and standardization constraints can ensure that the projected data remains independent, reduce information redundancy, and help maintain the scale consistency of the data, which is convenient for subsequent analysis. This constraint mechanism can make the mapping process more stable and reliable, thereby improving the performance of the entire hierarchical quantitative time series analysis method.

[0036] In practical applications, this constraint mechanism can be implemented through mathematical optimization methods. For example, orthogonality and normalization constraints can be used as constraints for the optimization problem, and the optimal projection matrix can be solved through an iterative optimization algorithm. The specific implementation method can be selected according to the needs of the actual problem and the situation of computing resources, such as gradient descent method, alternating direction multiplier method (ADMM), etc.

[0037] Furthermore, in some of the embodiments, the step of performing analysis according to the consistency representation to obtain an analysis result includes: According to the consistency representation, the total node time series data and the sub-node time series data, the relationship structure of the total node time series data and the sub-node time series data is obtained; The analysis results are obtained based on the consistency representation and relationship structure.

[0038] The technical solution of this application introduces a relational structure to link the originally independent consistent representation with the original data, so that the analysis process can simultaneously consider the unified representation and internal association of the data, thereby improving the depth and accuracy of the analysis. Compared with directly using consistent representation for analysis, this method can better retain and utilize the structural information of multi-level data, effectively improving the effect of hierarchical quantitative time series analysis.

[0039] Specifically, the technical solution of this application includes the following key features: Consistent representation: A unified representation obtained by mapping the time series data of the total node and sub-nodes into the same space. This can be achieved through various dimensionality reduction techniques, such as principal component analysis (PCA) or autoencoders.

[0040] Total node time series data and sub-node time series data: raw data of nodes at different levels in the system. These data can be collected by various sensors or measuring devices and may contain information at different time scales.

[0041] Relationship structure: obtained by analyzing the consistency representation and the original data, reflecting the association relationship between the total node and the sub-nodes. This can be achieved by constructing a graph structure or calculating a correlation coefficient matrix.

[0042] Analysis results: The final output of in-depth analysis based on consistent representation and relational structure. This can include anomaly detection, trend prediction, or pattern recognition.

[0043] In some embodiments, the analysis result may refer to a mathematical model obtained by analyzing the prepared input data, ie, the total node time series data and the sub-node time series data through the above method.

[0044] In some other embodiments, the analysis result may refer to a result such as trend prediction, anomaly detection or pattern recognition obtained after the mathematical model is obtained and the input data is input into the mathematical model.

[0045] The association and interaction between the above features are reflected in the following aspects: First, the consistent representation provides a unified representation space for the total node time series data and the sub-node time series data, so that data at different levels can be directly compared and analyzed. Second, the introduction of the relational structure utilizes the hierarchical information in the original data and supplements the details that may be ignored by the consistent representation. Finally, in the final analysis, the consistent representation and the relational structure are considered simultaneously, realizing a comprehensive analysis of multiple dimensions and scales.

[0046] In this way, the technical solution of the present application has the following advantages compared with the method of only using consistent representation for analysis: More original data information, especially hierarchical structure information, is retained, avoiding possible information loss during dimensionality reduction.

[0047] It can capture the complex relationships between nodes at different levels, which helps to discover deeper data patterns and regularities.

[0048] The robustness of the analysis is improved because multiple data representations are considered simultaneously, reducing the bias that may be introduced by a single representation.

[0049] The interpretability of analysis results is enhanced because the relationship structure provides additional contextual information, which helps to understand the causes of the analysis results.

[0050] Further, in some of the embodiments, the step of obtaining the relationship structure of the total node time series data and the child node time series data according to the consistency representation, the total node time series data and the child node time series data includes: Construct a graph structure based on the total node time series data and the sub-node time series data to obtain a relationship structure; Also includes: Regularize the graph structure to obtain a regularized term of the graph structure; Regularizing the consistent representation to obtain a regularization term of the consistent representation; The relational structure is obtained according to the graph structure, the regularization term of the graph structure and the regularization term of the consistency representation.

[0051] Among them, constructing a graph structure means that by converting the total node time series data and the sub-node time series data into a graph structure, the relationship between nodes can be intuitively represented. The graph structure can be represented in the form of an adjacency matrix or an adjacency list.

[0052] Regularization refers to regularizing the graph structure and consistent representation, which can reduce overfitting and improve the generalization ability of the model.

[0053] By comprehensively considering the graph structure, the regularization term of the graph structure, and the regularization term of the consistent representation, a more accurate relational structure can be obtained.

[0054] The technical solution of the present application can more intuitively represent the relationship between nodes by constructing a graph structure and introducing regularization processing. The introduction of regularization processing can more accurately quantify the degree of association between nodes and improve the generalization ability of the model. Ultimately, by comprehensively considering the graph structure and regularization terms, a more accurate and reliable relationship structure is obtained.

[0055] Through this method, the present application can more accurately capture the complex relationship between the total node and the sub-nodes, and improve the accuracy and reliability of time series analysis. Compared with the traditional method, the technical solution of the present application takes into account the multi-level data structure and improves the robustness of the model through regularization processing, thereby showing better performance when processing complex hierarchical measurement time series data.

[0056] Specifically, in some of the embodiments, the formula for mapping the total node time series data and the sub-node time series data to the same space to obtain a consistent representation is: ; in, Indicates The data matrix input from each perspective includes the total node perspective and the sub-node perspective, corresponding to the total node time series data and the sub-node time series data; Indicates the summary point perspective; Indicates the child node perspective; Indicates The projection matrix of the view input; Indicates consistency representation; represents the regularization parameter; is the square of the Frobenius norm; The orthogonality and normalization constraints are: ; in, and Both represent the identity matrix; represents the dimension of the identity matrix; Represents a transpose operation.

[0057] In some specific embodiments, the method provided in the present application can be used to analyze power consumption data.

[0058] in, ∈ , represents the power consumption data recorded from a macro perspective, that is, the total node perspective. represents the number of days, and the number of rows corresponds to the 24 hours in a day. represents the field of real numbers.

[0059] in, ∈ , as a description The second perspective of the electricity consumption of the day, is the number of nodes, so the number of rows of node perspective data is 24 This perspective is similar to In comparison, detailed power consumption data for each sub-node is provided, and the comprehensiveness of the information helps in the subsequent cluster analysis, including identifying the power consumption patterns of components.

[0060] is a projection-based consensus,representation learning model, where, ∈ , Indicates the corresponding The number of rows of electricity consumption data in each perspective is calculated by and Project to the same space and extract the consistent representation of the total node perspective and the sub-node perspective , in order to avoid trivial solutions where both the projection and the consistency representation are 0, we introduce As orthogonality and normalization constraints, used to constrain the projection.

[0061] Specifically, the formula for the relational structure is obtained based on the graph structure, the regularization term of the graph structure, and the regularization term of the consistency representation: ; in, and It represents different time segments; It indicates the total number of time segments; It represents the similarity matrix; It means for The regularization parameter of ; represents the transpose operation, that is, It means that The transpose of It means for The regularization parameter of ; It represents the trace operation; is the square of the Euclidean norm; It also includes constraints, which are: .

[0062] In this formula, and Represents the total node perspective and the sub-node perspective The data of the column, that is, the total node component and the sub-node component Electricity consumption in hours per day.

[0063] in, ∈ Reflects The distribution structure of daily electricity consumption data provides support for clustering tasks. Indicates Daily electricity consumption and The similarity between the electricity consumption of the day. In the alternating optimization process, and Promote each other. is the graph regularization parameter, where is the graph Laplacian matrix, , ∈ , is a diagonal matrix, whose The diagonal elements are Calculate, where is a matrix diagonal elements of .

[0064] Specifically, the steps of performing analysis according to the consistency representation to obtain the analysis result include: Construct mathematical models based on consistent representations; Analyze according to mathematical model to obtain analysis results; The mathematical model is: ; , ; The steps of obtaining the analysis results according to the mathematical model include: from , , In the process of changing one parameter in turn and fixing other parameters for alternate optimization, it includes: fixed , ,renew ; fixed , ,renew ; fixed , ,renew .

[0065] Specifically, for the above mathematical model, this application adopts an alternating optimization strategy, optimizing only one variable and fixing other variables alternately, to calculate the local optimal solution of the model. Specifically, this application decomposes the original optimization problem into three sub-problems, corresponding to the three variables to be optimized: , , .

[0066] (1) When fixed , ,renew When , the optimization objective problem can be simplified to: ; In order to find the minimum value of the above formula, we can take its partial derivative and make the result zero, and we can get: ; By transforming the above formula, we can get the optimal as follows: .

[0067] (2) When fixed , ,renew When , the optimization objective problem can be simplified to: ; ; Expanding the Frobenius norm into a trace form, the optimization objective problem can be further simplified as follows: ; ; Continuing to sort out the first and fourth traces of the above formula, we can get: ; =

[0068] =

[0069] The following formula can be rewritten: ; ; get: ; ; The above formula can be transformed: ; ; make and , the above formula can be simplified to: ; , ; The above formula is a typical orthogonal Proust problem, which can be calculated by singular value decomposition and has The closed-form solution of Finally it can be calculated as follows: = ; in, and They are The left singular vector matrix and the right singular vector matrix of the singular value decomposition of .

[0070] (3) When fixed , ,renew When , the optimization objective problem can be simplified to: ; ; Transform the above formula: ; =

[0071] =

[0072] =

[0073] ; in, , independent of , the above formula can be simplified to: ; ; According to the TTK condition, the closed-form solution of the above formula can be obtained as: ; in, is the number of nearest neighbors; It is a traversal index, which is used as a temporary variable to mark the currently accumulated node.

[0074] Finally, for the above mathematical model: ; , ; This can be done by: ; = ; ; These three equations are solved, where solving means obtaining the , , The optimal values ​​of these three variables.

[0075] Second, refer to Figure 2 The present application also proposes a hierarchical measurement time series analysis device for performing time series analysis on a system composed of multiple sub-nodes, including: The acquisition module 210 is used to acquire the total node time series data and the sub-node time series data of the system; A mapping module 220, used to map the total node time series data and the sub-node time series data to the same space to obtain a consistent representation; The analysis module 230 is used to perform analysis according to the consistency representation to obtain an analysis result.

[0076] Through the above technical solution, the time series data of the total node and the child nodes are mapped to the same space and analyzed based on the consistent representation, which makes full use of the multi-level data information, thereby solving the problem of single-perspective analysis in the prior art that data details are ignored. It has the advantage of being able to make full use of multi-level data information and improve the accuracy and effect of clustering analysis.

[0077] In addition, in some preferred embodiments, a hierarchical metrology time series analysis device proposed in the present application can perform any step of the above method.

[0078] Thirdly, refer to Figure 3The present application also provides an electronic device, including a processor 310 and a memory 320, wherein the memory 320 stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the steps in the above method are executed.

[0079] Through the above technical solution, the processor 310 and the memory 320 are interconnected and communicate with each other through a communication bus and / or other forms of connection mechanisms (not marked), and the memory 320 stores computer-readable instructions executable by the processor 310. When the electronic device is running, the processor 310 executes the computer-readable instructions to execute the method in any optional implementation method of the above embodiment to achieve the following functions: obtain the system's total node time series data and sub-node time series data; map the total node time series data and the sub-node time series data to the same space to obtain a consistent representation; and perform analysis based on the consistent representation to obtain an analysis result.

[0080] In a fourth aspect, the present application also provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps in the above method are executed.

[0081] Through the above technical solution, when the computer program is executed by the processor, the method in any optional implementation method of the above embodiment is executed to achieve the following functions: obtaining the total node time series data and sub-node time series data of the system; mapping the total node time series data and the sub-node time series data to the same space to obtain a consistent representation; and performing analysis based on the consistent representation to obtain an analysis result.

[0082] Among them, the computer-readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable red-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0083] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0084] In addition, the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0085] Furthermore, the functional modules in the various embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

[0086] The above description is only an embodiment of the present application and is not intended to limit the protection scope of the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0087] The above description is only an embodiment of the present application and is not intended to limit the protection scope of the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A hierarchical quantitative time series analysis method for performing time series analysis on a system consisting of multiple sub-nodes, characterized in that: include: Obtaining total node time series data and sub-node time series data of the system; Mapping the total node time series data and the sub-node time series data to the same space to obtain a consistent representation; An analysis result is obtained by performing analysis according to the consistency expression.

2. A hierarchical quantitative time series analysis method according to claim 1, characterized in that: The method also includes the steps of: The projection matrix in the process of mapping the total node time series data and the sub-node time series data into the same space to obtain a consistent representation is constrained by orthogonality and standardization constraints.

3. A hierarchical quantitative time series analysis method according to claim 2, characterized in that: The step of analyzing according to the consistency representation to obtain an analysis result comprises: According to the consistency representation, the total node time series data and the sub-node time series data, acquiring a relationship structure between the total node time series data and the sub-node time series data; An analysis result is obtained by performing analysis according to the consistency representation and the relationship structure.

4. A hierarchical quantitative time series analysis method according to claim 3, characterized in that: The step of acquiring the relationship structure of the total node time series data and the sub-node time series data according to the consistency representation, the total node time series data and the sub-node time series data comprises: Constructing a graph structure according to the total node time series data and the sub-node time series data to obtain the relationship structure; Also includes: Performing regularization processing on the graph structure to obtain a regularization term of the graph structure; Performing regularization processing on the consistency representation to obtain a regularization term of the consistency representation; The relational structure is obtained according to the graph structure, the regularization term of the graph structure, and the regularization term of the consistency representation.

5. A hierarchical quantitative time series analysis method according to claim 4, characterized in that: The formula for mapping the total node time series data and the sub-node time series data to the same space to obtain a consistent representation is: ； in, Indicates The data matrix input from each perspective includes the total node perspective and the sub-node perspective, corresponding to the total node time series data and the sub-node time series data; Indicates the summary point perspective; Indicates the child node perspective; Indicates The projection matrix of the view input; Indicates consistency representation; represents the regularization parameter; is the square of the Frobenius norm; The orthogonality and normalization constraints are: ； in, and Both represent the identity matrix; represents the dimension of the identity matrix; Represents a transpose operation.

6. A hierarchical quantitative time series analysis method according to claim 5, characterized in that: The formula for obtaining the relational structure according to the graph structure, the regularization term of the graph structure and the regularization term of the consistency representation is: ； in, and It represents different time segments; It indicates the total number of time segments; It represents the similarity matrix; It means for The regularization parameter of ; represents the transpose operation, that is, It means that The transpose of It means for The regularization parameter of ; It represents the trace operation; is the square of the Euclidean norm; It also includes constraints, which are: 。 7. A hierarchical quantitative time series analysis method according to claim 6, characterized in that: The step of analyzing according to the consistency representation to obtain an analysis result comprises: constructing a mathematical model based on the consistency representation; Performing analysis according to the mathematical model to obtain analysis results; The mathematical model is: ； , ； The step of analyzing according to the mathematical model to obtain the analysis result comprises: from , , In the process of changing one parameter in turn and fixing other parameters for alternate optimization, it includes: fixed , ,renew ; fixed , ,renew ; fixed , ,renew .

8. A hierarchical measurement time series analysis device, used for performing time series analysis on a system consisting of multiple sub-nodes, characterized in that: include: An acquisition module, used to acquire the total node time series data and the sub-node time series data of the system; A mapping module, used for mapping the total node time series data and the sub-node time series data to the same space to obtain a consistent representation; An analysis module is used to perform analysis according to the consistency representation to obtain an analysis result.

9. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the steps in the method according to any one of claims 1 to 7 are executed.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are executed.

Images