Carbon emission monitoring user classification method and system, storage medium and equipment

Abstract

The invention provides a carbon emission monitoring user classification method, which is characterized by comprising the following steps of: acquiring energy consumption data of carbon emission monitoring users, and checking vacant data of the energy consumption data; performing normalization processing on the checked and complemented energy consumption data; constructing an FCM clustering model based on a quantum flower pollination optimization algorithm; organizing the normalized energy consumption data by using a date-based method, and clustering the normalized energy consumption data by using the FCM clustering model; and carrying out dimensionality reduction on the clustered energy consumption data by using a t-SNE dimensionality reduction method, and visualizing a clustering result. According to the method, clustering processing is firstly carried out on the data, then a t-SNE big data dimension reduction method is adopted to reduce the complexity of energy consumption data, the analysis efficiency is improved, meanwhile, key information is reserved, the clustering result is more accurate and reliable and can be visualized, the relation between the data is fully analyzed, and errors are reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of carbon emission monitoring, and in particular relates to a carbon emission monitoring user classification method, system, storage medium and equipment. Background Art

[0002] In the context of addressing climate change and promoting green development, accurate monitoring and management of carbon emissions have become particularly important. With the development of big data technology, it has become possible to conduct in-depth analysis of various energy consumption data of key carbon emission monitoring users. Cluster analysis is an unsupervised learning method that divides users into different categories based on the similarity between data points, helping to identify user groups with similar energy consumption characteristics. This method can reveal the differences in energy consumption patterns between different users and provide a scientific basis for formulating differentiated emission reduction strategies. At the same time, big data dimensionality reduction technology reduces the complexity of analysis and improves computing efficiency by reducing the dimensions of the data.

[0003] In the existing technology, only the user's electricity consumption data is counted, and the heat consumption data, gas consumption data, and various energy consumption data such as coal, coke, crude oil, gasoline, etc. are not counted; the results of cluster analysis are often affected by the initial value, and different initial values ​​may lead to different clustering results; cluster analysis also needs to determine the number of categories in advance, but this number is often difficult to determine, and different numbers of categories may lead to different clustering results; cluster analysis usually assumes that the data is linearly separable, but for nonlinear data, the clustering effect may not be good. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides a method for classifying users of carbon emission monitoring, the method comprising:

[0005] Collect energy consumption data of carbon emission monitoring users, and check and supplement missing data in the energy consumption data;

[0006] Normalize the energy consumption data after checking and supplementing;

[0007] Construct an FCM clustering model based on the quantum flower pollination optimization algorithm;

[0008] Organizing the normalized energy consumption data using a date-based method, and clustering the normalized energy consumption data using the FCM clustering model;

[0009] The t-SNE dimensionality reduction method is used to reduce the dimension of the clustered energy consumption data and visualize the clustering results.

[0010] Furthermore, the energy consumption data includes electricity consumption data, heat consumption data, gas consumption data, and coal, coke, crude oil, gasoline, kerosene, diesel and fuel oil consumption data.

[0011] Furthermore, the heat usage data is obtained through a heat meter;

[0012] The gas usage data is obtained through a gas meter;

[0013] The coal, coke, crude oil, gasoline, kerosene, diesel and fuel oil consumption data are obtained by counting the consumption of users;

[0014] The power consumption data is obtained through the power meter

[0015] Furthermore, the checking and supplementing of missing data includes:

[0016] Improving Generative Adversarial Networks;

[0017] Extracting feature data of the energy consumption data, and using the improved generative adversarial network to perform deep learning using the feature data;

[0018] Use the generative adversarial network after deep learning to complete the energy consumption data.

[0019] Furthermore, the improved generative adversarial network comprises:

[0020] Build a GAN model, including a generator network and a discriminator network;

[0021] Add a two-layer LSTM network to the generation network to obtain the generator;

[0022] The discriminant network is constructed as a convolutional neural network to obtain a discriminator;

[0023] The generator and the discriminator are used to conduct confrontation through alternating training.

[0024] Furthermore, the normalization processing includes: using an adaptive normalization method to dynamically adjust normalization parameters according to local characteristics of time series data and correlations between adjacent time series.

[0025] Furthermore, the construction of the FCM clustering model based on the quantum flower pollination optimization algorithm includes:

[0026] The cluster center of the normalized energy consumption data is regarded as pollen, and a flower pollination algorithm is constructed;

[0027] The quantized particle search mechanism is introduced into the flower pollination algorithm to obtain the quantized flower pollination optimization algorithm.

[0028] Furthermore, the flower pollination algorithm has two stages, including global pollination and local pollination.

[0029] Furthermore,

[0030] The global pollination stage is to simulate the long-distance spread of pollen to ensure the reproduction of flowers. The formula is:

[0031]

[0032] in, represents the cluster center of the user energy consumption data solved at the tth iteration; G * The optimal solution for the current group; is the proportional factor for controlling the moving step length; b1 is the error coefficient; L(λ) is the step size;

[0033] The local pollination is carried out by close contact of flowers under the influence of various natural factors, and the formula is expressed as:

[0034]

[0035] in, and represents any two different energy consumption data points in the same class; b2 is the error coefficient; δ∈[0,1] represents a random number that obeys a uniform distribution.

[0036] Furthermore, the step size is subject to the levy flight step size, which is calculated as follows:

[0037]

[0038] Where Γ(λ) represents the standard gamma function, s represents the step size, and λ is the distribution factor.

[0039] Furthermore, the quantized particle search mechanism is established, and its formula is expressed as:

[0040]

[0041] Where t is the number of iterations; is the attracting particle that exists in the jth dimension for any particle i; is the position of particle i in the jth dimension at the t+1th generation; η is the shrinkage and expansion factor; represents the average position of the group; is a random number uniformly distributed on (0,1).

[0042] Furthermore, the quantized flower pollination optimization algorithm includes: introducing a quantized particle search mechanism into the flower pollination algorithm to obtain a quantized flower pollination optimization algorithm, the formula of which is expressed as follows:

[0043]

[0044] Furthermore, the calculation formula of the objective function value of the FCM algorithm is expressed as:

[0045]

[0046] Among them, u represents the degree of membership, x i represents a particle, i.e., the energy consumption data of key carbon emission monitoring users, c represents the cluster center, and m is the fuzzy index;

[0047] The membership degree formula is expressed as:

[0048]

[0049] The formula of the cluster center is expressed as,

[0050]

[0051] Furthermore,

[0052] The iterative optimization includes: outputting the optimal solution until an iteration stop condition is met;

[0053] The iteration stopping condition is that the iteration number reaches a preset value or the improvement of the objective function is less than a preset threshold, and the iteration is stopped.

[0054] Furthermore, the dimensionality reduction of the clustered user energy consumption data using the t-SNE dimensionality reduction method includes:

[0055] Calculate the probability distribution of data points in high-dimensional space and calculate the probability distribution of data points in low-dimensional space;

[0056] Use KL divergence to represent the difference in probability distribution between low-dimensional space and high-dimensional space;

[0057] Use the gradient descent method to optimize the KL divergence, obtain the dimensionality reduction result, and visualize the dimensionality reduction result.

[0058] Furthermore, the probability distribution between two high-dimensional data points is calculated, and its formula is expressed as:

[0059]

[0060] Among them, p i|j Indicates that the center is m i When, m j is the probability of its neighboring points; p i|j Indicates that the center is m j When, m i is the probability of its neighboring points; N is N high-dimensional energy consumption data {m1,m2,…,m N};

[0061] The probability distribution of the data points in the low-dimensional space is calculated using a t distribution with a degree of freedom of 1, and its formula is expressed as follows:

[0062]

[0063] Among them, the data points in the low-dimensional space are represented as {n1,n2,…,n N};

[0064] The KL divergence is used to represent the difference in probability distribution between low-dimensional space and high-dimensional space, and its formula is expressed as follows:

[0065]

[0066] The gradient descent method is used to optimize the KL divergence, and its formula is expressed as follows:

[0067]

[0068] Furthermore, the visualizing the dimensionality reduction result includes: generating a clustering result diagram for each category according to the dimensionality reduction result, and visualizing the clustering result.

[0069] A carbon emission monitoring user classification system, the system comprising: a data acquisition module, a data processing module, a clustering model construction module, a data clustering module and a data dimension reduction module;

[0070] The data collection module is used to collect various energy consumption data of key carbon emission detection users;

[0071] The data processing module is used to normalize various energy consumption data of key carbon emission monitoring users and fill in the missing data of the original data;

[0072] The clustering model construction module is used to construct an FCM clustering model based on a quantum flower pollination optimization algorithm;

[0073] The data clustering module is used to organize data based on the date method using the FCM clustering model constructed by the clustering model construction module, and cluster the energy data of key carbon emission monitoring users;

[0074] The data dimension reduction module uses the t-SNE dimension reduction method to reduce the dimension of the clustered user energy consumption data and visualize the clustering results.

[0075] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, any of the method steps described above is implemented.

[0076] An electronic device comprises a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0077] Memory, used to store computer programs;

[0078] The processor is used to implement any of the above-mentioned method steps when executing the program stored in the memory.

[0079] Compared with the prior art, the present invention has the following advantages:

[0080] 1. The present invention is different from the existing method which only counts the electricity consumption data of users. It also counts the heat consumption data, gas consumption data and various energy consumption data such as coal, coke, crude oil, gasoline, etc. for comprehensive analysis. The FCM clustering method optimized by the quantum flower pollination optimization algorithm is used to analyze the energy consumption data of key carbon emission monitoring users. The clustering center of the user's energy consumption data is regarded as pollen for optimization, which improves the clustering effect, determines the energy usage pattern and trend of the user, and classifies users with similar energy consumption patterns into the same category, revealing the laws and characteristics of user energy usage.

[0081] 2. The present invention first performs clustering processing on the data, and then uses the t-SNE big data dimensionality reduction method to reduce the complexity of the energy consumption data, thereby improving the analysis efficiency while retaining key information, making the clustering results more accurate, reliable and visualized, fully analyzing the relationship between the data and reducing errors.

[0082] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0084] Figure 1 A flow chart of a carbon emission monitoring user classification method is shown.

[0085] Figure 2 A schematic diagram of a carbon emission monitoring user classification system module is shown.

[0086] Figure 3 A schematic diagram of the clustering results of the first category of key carbon emission monitoring users is shown.

[0087] Figure 4 A schematic diagram of the clustering results of the second category of key carbon emission monitoring users is shown.

[0088] Figure 5 A schematic diagram of the clustering results of the third category of key carbon emission monitoring users is shown.

[0089] Figure 6 A schematic diagram of the clustering results of the fourth category of key carbon emission detection users is shown. DETAILED DESCRIPTION

[0090] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0091] like Figure 1 As shown, a method for classifying carbon emission monitoring users comprises the following steps:

[0092] S1. Collect various energy consumption data of key carbon emission monitoring users, and check and supplement the missing data in the collected energy consumption data.

[0093] S1.1. The energy consumption data collected include electricity consumption data (provided by power companies), heat consumption data (provided by urban heating companies), gas consumption data (provided by urban gas companies or directly connected to natural gas), and data on direct use of coal, coke, crude oil, gasoline, kerosene, diesel and fuel oil by special industries.

[0094] Optionally, the thermal data is obtained by using a thermal meter;

[0095] Gas consumption data is obtained by using a gas meter;

[0096] Data on coal, coke, crude oil, gasoline, kerosene, diesel and fuel oil are obtained by counting the usage of users;

[0097] In order to improve the calculation accuracy of electricity consumption data, an ultra-fine granularity smart electricity meter is installed to collect electricity consumption data in one minute to obtain user usage data.

[0098] S1.2. Use improved generative adversarial networks (GANs) to complete the missing energy consumption data.

[0099] S1.2.1. Improve generative adversarial networks (GANs).

[0100] The GAN model is constructed, including a generator network and a discriminator network. Since LSTM has a good ability to maintain timing, a double-layer LSTM network is added to the generator network, and the discriminator network is constructed as a convolutional neural network (CNN).

[0101] S1.2.2. Extract the characteristic data of the collected energy consumption data, conduct deep learning on these characteristic data, and use the generator and discriminator to conduct confrontation through alternating training.

[0102] S1.2.3. After training is completed, use the generator to complete the missing energy consumption data.

[0103] Optionally, the generator generates filling values ​​similar to the real data distribution based on the existing data and the context information of the actual data, thereby completing the missing energy consumption data of the user.

[0104] S2. Normalize various energy consumption data of key carbon emission monitoring users.

[0105] S2.1. Adaptive normalization method is used to dynamically adjust normalization parameters according to the local characteristics of time series data and the correlation between adjacent time series.

[0106] S2.1.1. The energy consumption data of the user is segmented according to the time window size to form multiple subsequences, and the statistical parameters of each subsequence are calculated.

[0107] Optionally, the statistical parameters include: mean, standard deviation, maximum value, minimum value, etc.

[0108] S2.1.2. Consider the correlation between adjacent subsequences and estimate the normalization parameters through sliding average to better reflect the dynamic changes of time series.

[0109] S2.1.3. Use the estimated normalization parameter to normalize each subsequence to obtain a normalized value, and output the normalized user energy consumption data for subsequent analysis.

[0110] Optionally, during the normalization process, the normalization parameters may be dynamically adjusted according to the normalization results and data characteristics to further improve the normalization effect.

[0111] S3. Construct an FCM clustering model based on the quantum flower pollination optimization algorithm.

[0112] S3.1. The cluster center of the normalized energy consumption data is regarded as pollen to implement the flower pollination algorithm, which mainly has two stages: global pollination and local pollination.

[0113] S3.1.1. In the global pollination stage, pollen is carried by insects and other spreaders, thereby achieving long-distance pollen transmission and ensuring flower reproduction. The formula is expressed as:

[0114]

[0115] in, is the solution of the i-th pollen when the iteration number is t, and also represents the cluster center of the user energy consumption data solved at the t-th iteration; G * The optimal solution for the current group; is the proportional factor that controls the moving step size; b1 is the error coefficient; L(λ) is the step size.

[0116] Optionally, L(λ) is subject to the levy flight step length, which is calculated as,

[0117]

[0118] Where Γ(λ) represents the standard gamma function, s represents the step size, and λ is the distribution factor.

[0119] S3.1.2. During the local pollination stage, there are various natural factors that bring flowers into close contact for pollination:

[0120]

[0121] in, and They represent any two pollens from different flowers of similar flowering plants, and also represent any two different energy consumption data points in the same class; b2 is the error coefficient; δ∈[0,1] represents a random number that obeys a uniform distribution, so that the local pollination process is a local random walk process.

[0122] S3.2. Establish a quantized particle search mechanism.

[0123] S3.2.1. A quantum system uses particles in quantum space to describe each individual. Assuming the number of iterations is t, any particle i has an attracting particle in the jth dimension. Then the quantized particle search mechanism is expressed as:

[0124]

[0125] in, is the position of particle i in the t+1th generation on the jth dimension; η is the contraction and expansion factor, which is used to adjust the convergence speed of the algorithm and changes linearly with the number of iterations; represents the average position of the group; is a random number uniformly distributed on (0,1).

[0126] S3.3. Assuming that the flower pollination system is also a quantum system, the quantized particle search mechanism is introduced into the flower pollination algorithm to obtain the attraction point

[0127] S3.3.1. In the flower pollination algorithm, it is assumed that the attractor Generated for the global pollination phase Then substitute the global pollination stage equation into the quantized particle search mechanism equation, and the formula is expressed as:

[0128]

[0129] S3.4. Use the objective function value of the FCM clustering model to calculate the membership degree and cluster center.

[0130] S3.4.1. The calculation formula of the objective function value of the FCM clustering model is expressed as:

[0131]

[0132] Among them, u represents the degree of membership, x i represents a particle, that is, the energy consumption data of key carbon emission monitoring users, c represents the cluster center, and m is the fuzzy index.

[0133] S3.4.2, the membership degree formula is expressed as:

[0134]

[0135] The formula of the cluster center is expressed as:

[0136]

[0137] S3.5. Repeat steps S3.1-S3.4, perform iterative optimization based on the quantum flower pollination optimization algorithm to output the optimal solution, and use the obtained optimal solution as the initial energy consumption data clustering center of the FCM clustering model to perform energy consumption clustering analysis on key carbon emission monitoring users.

[0138] Optionally, the iterative optimization includes: outputting the optimal solution until an iteration stopping condition is met.

[0139] Optionally, the iteration stopping condition is that the iteration is stopped when the number of iterations reaches a preset value or the improvement of the objective function is less than a certain threshold.

[0140] S4. Use a date-based method to organize the normalized energy consumption data and cluster the energy consumption data of key carbon emission monitoring users.

[0141] S4.1. Start clustering from the start and end date range of the user data being counted. In this group of days, each user will generate an average energy consumption data.

[0142] S4.2. According to the industry categories to which the statistical users belong, through investigation and analysis of the typical energy consumption patterns of each industry, based on these experiences and prior knowledge, estimate a reasonable number of clusters.

[0143] S4.3. Use the FCM clustering model optimized by the quantum flower pollination optimization algorithm to process user energy consumption data, and continuously adjust the number of clusters based on the elbow rule to classify key carbon emission monitoring users.

[0144] S5. Use the t-SNE dimensionality reduction method to reduce the dimension of the clustered user energy consumption data and visualize the clustering results.

[0145] S5.1. Use t-SNE to reduce the clustered energy consumption data to 2 dimensions.

[0146] S5.1.1 For N high-dimensional energy consumption data {m1,m2,…,m N}, the probability distribution between high-dimensional data points is calculated as:

[0147]

[0148] Among them, when p i|j Indicates that the center is m i When, m j As the probability of its neighboring points, Gaussian distribution is adopted, and its formula is expressed as:

[0149]

[0150] S5.1.2、The data points in the low-dimensional space are {n1,n2,…,n N}, using the t distribution with a degree of freedom of 1, its probability distribution is expressed as:

[0151]

[0152] The difference between the probability distribution of low-dimensional space and the probability distribution of high-dimensional space is expressed by KL divergence:

[0153]

[0154] The smaller the C value, the higher the simulation accuracy. The gradient descent method is used to optimize the KL divergence, and its formula is expressed as:

[0155]

[0156] S5.2, such as Figure 3-Figure 6As shown in the figure, based on the dimensionality reduction results, a clustering result diagram for each category is generated to visualize the clustering results.

[0157] like Figure 2 As shown, a carbon emission monitoring user classification system includes: a data acquisition module, a data processing module, a clustering model construction module, a data clustering module and a data dimension reduction module.

[0158] The data collection module is used to collect various energy consumption data of key carbon emission detection users.

[0159] The data processing module is used to normalize various energy consumption data of key carbon emission monitoring users and fill in the missing data of the original data.

[0160] The clustering model construction module is used to construct an FCM clustering model based on the quantum flower pollination optimization algorithm.

[0161] The data clustering module is used to use the FCM clustering model constructed by the clustering model construction module to organize data based on the date method and cluster the energy data of key carbon emission monitoring users.

[0162] The data dimension reduction module uses the t-SNE dimension reduction method to reduce the dimension of the clustered user energy consumption data and visualize the clustering results.

[0163] Based on the above disclosed content, the present invention also provides an electronic device accordingly. The electronic device of an embodiment of the present invention includes at least one electrically connected processor and at least one storage medium, the storage medium is electrically connected to the processor, wherein the storage medium stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the method as described above.

[0164] Based on the same inventive concept, the present invention also provides a storage medium, which stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method as described above.

[0165] The above description and the accompanying drawings fully illustrate the embodiments of the present invention so that those skilled in the art can practice them. Other embodiments may include structural and other changes. The embodiments represent only possible variations. Unless explicitly required, individual components and functions are optional, and the order of operations may vary. Parts and features of some embodiments may be included in or replace parts and features of other embodiments. The embodiments of the present invention are not limited to the structures described above and shown in the accompanying drawings, and various modifications and changes may be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. A method for classifying users of carbon emission monitoring, characterized in that: The method comprises, Collect energy consumption data of carbon emission monitoring users, and check and supplement missing data in the energy consumption data; Normalize the energy consumption data after checking and supplementing; Construct an FCM clustering model based on the quantum flower pollination optimization algorithm; Organizing the normalized energy consumption data using a date-based method and clustering it using the FCM clustering model; The t-SNE dimensionality reduction method is used to reduce the dimension of the clustered energy consumption data and visualize the clustering results.

2. According to claim 1, the carbon emission monitoring user classification method is characterized in that: The energy consumption data includes electricity consumption data, heat consumption data, gas consumption data, and coal, coke, crude oil, gasoline, kerosene, diesel and fuel oil consumption data.

3. The carbon emission monitoring user classification method according to claim 2 is characterized in that: The heat consumption data is obtained through a heat meter; The gas usage data is obtained through a gas meter; The coal, coke, crude oil, gasoline, kerosene, diesel and fuel oil consumption data are obtained by counting the consumption of users; The electricity consumption data is obtained through an electric energy meter.

4. The carbon emission monitoring user classification method according to claim 1 is characterized in that: The said checking and supplementing of missing data includes: Improving Generative Adversarial Networks; Extracting feature data of the energy consumption data, and using the improved generative adversarial network to perform deep learning using the feature data; Use the generative adversarial network after deep learning to complete the energy consumption data.

5. The carbon emission monitoring user classification method according to claim 4 is characterized in that: The improved generation adversarial network comprises: Build a GAN model, including a generator network and a discriminator network; Add a two-layer LSTM network to the generation network to obtain the generator; The discriminant network is constructed as a convolutional neural network to obtain a discriminator; The generator and the discriminator are used to conduct confrontation through alternating training.

6. The carbon emission monitoring user classification method according to claim 1 is characterized in that: The normalization process includes: using an adaptive normalization method to dynamically adjust normalization parameters according to local characteristics of time series data and correlations between adjacent time series.

7. The carbon emission monitoring user classification method according to claim 1 is characterized in that: The construction of the FCM clustering model based on the quantum flower pollination optimization algorithm includes: The cluster center of the normalized energy consumption data is regarded as pollen, and a flower pollination algorithm is constructed; The quantized particle search mechanism is introduced into the flower pollination algorithm to obtain the quantized flower pollination optimization algorithm.

8. The carbon emission monitoring user classification method according to claim 7 is characterized in that: The flower pollination algorithm has two stages, including global pollination and local pollination.

9. The carbon emission detection user classification method according to claim 8 is characterized in that: The global pollination stage is to simulate the long-distance spread of pollen to ensure the reproduction of flowers. The formula is: in, represents the cluster center of the user energy consumption data solved at the tth iteration; G * The optimal solution for the current group; is the proportional factor for controlling the moving step length; b1 is the error coefficient; L(λ) is the step size; The local pollination is carried out by close contact of flowers under the influence of various natural factors, and the formula is expressed as: Among them, Y j t and represents any two different energy consumption data points in the same class; b2 is the error coefficient; δ∈[0,1] represents a random number that obeys a uniform distribution.

10. The carbon emission monitoring user classification method according to claim 9 is characterized in that: The step size is subject to the levy flight step size, which is calculated as follows: Where Γ(λ) represents the standard gamma function, s represents the step size, and λ is the distribution factor.

11. The carbon emission monitoring user classification method according to claim 7, characterized in that: The quantized particle search mechanism is established, and its formula is expressed as: Where t is the number of iterations; is the attracting particle that exists in the jth dimension for any particle i; is the position of particle i in the jth dimension at the t+1th generation; η is the shrinkage and expansion factor; represents the average position of the group; is a random number uniformly distributed on (0,1).

12. The carbon emission monitoring user classification method according to claim 7 is characterized in that: The quantized flower pollination optimization algorithm includes: introducing a quantized particle search mechanism into the flower pollination algorithm to obtain a quantized flower pollination optimization algorithm, the formula of which is expressed as follows:

13. The carbon emission monitoring user classification method according to claim 7 is characterized in that: The calculation formula of the objective function value of the FCM algorithm is expressed as: Among them, u represents the degree of membership, x i represents a particle, i.e., the energy consumption data of key carbon emission monitoring users, c represents the cluster center, and m is the fuzzy index; The membership degree formula is expressed as: The formula of the cluster center is expressed as, 14. The carbon emission monitoring user classification method according to claim 7 is characterized in that: The iterative optimization includes: outputting the optimal solution until an iteration stop condition is met; The iteration stopping condition is that the iteration number reaches a preset value or the improvement of the objective function is less than a preset threshold, and the iteration is stopped.

15. The carbon emission monitoring user classification method according to claim 1 is characterized in that: The dimensionality reduction method of using the t-SNE to reduce the dimension of the clustered user energy consumption data includes: Calculate the probability distribution of data points in high-dimensional space and calculate the probability distribution of data points in low-dimensional space; Use KL divergence to represent the difference in probability distribution between low-dimensional space and high-dimensional space; Use the gradient descent method to optimize the KL divergence, obtain the dimensionality reduction result, and visualize the dimensionality reduction result.

16. The carbon emission monitoring user classification method according to claim 15, characterized in that: The formula for calculating the probability distribution between two high-dimensional data points is as follows: Among them, p i|j Indicates that the center is m i When, m j is the probability of its neighboring points; p j|i Indicates that the center is m j When, m i is the probability of its neighboring points; N is N high-dimensional energy consumption data {m1,m2,…,m N }; The probability distribution of the data points in the low-dimensional space is calculated using a t distribution with a degree of freedom of 1, and its formula is expressed as follows: Among them, the data points in the low-dimensional space are represented as {n1,n2,…,n N }; The KL divergence is used to represent the difference in probability distribution between low-dimensional space and high-dimensional space, and its formula is expressed as follows: The gradient descent method is used to optimize the KL divergence, and its formula is expressed as follows:

17. The carbon emission monitoring user classification method according to claim 16, characterized in that: The visualizing the dimensionality reduction result includes: generating a clustering result diagram for each category according to the dimensionality reduction result, and visualizing the clustering result.

18. A carbon emission monitoring user classification system, the system comprising: Data acquisition module, data processing module, clustering model construction module, data clustering module and data dimension reduction module; The data collection module is used to collect various energy consumption data of key carbon emission detection users; The data processing module is used to normalize various energy consumption data of key carbon emission monitoring users and fill in the missing data of the original data; The clustering model construction module is used to construct an FCM clustering model based on a quantum flower pollination optimization algorithm; The data clustering module is used to organize data based on the date method using the FCM clustering model constructed by the clustering model construction module, and cluster the energy data of key carbon emission monitoring users; The data dimension reduction module uses the t-SNE dimension reduction method to reduce the dimension of the clustered user energy consumption data and visualize the clustering results.

19. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method steps described in any one of claims 1 to 16 are implemented.

20. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus; Memory, used to store computer programs; A processor, for implementing the method steps described in any one of claims 1-16 when executing a program stored in a memory.

Images