CN113591947A - Power data clustering method and device based on power consumption behaviors and storage medium - Google Patents
Power data clustering method and device based on power consumption behaviors and storage medium Download PDFInfo
- Publication number
- CN113591947A CN113591947A CN202110804808.5A CN202110804808A CN113591947A CN 113591947 A CN113591947 A CN 113591947A CN 202110804808 A CN202110804808 A CN 202110804808A CN 113591947 A CN113591947 A CN 113591947A
- Authority
- CN
- China
- Prior art keywords
- user
- evaluation index
- ideal solution
- obtaining
- acquiring
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The application discloses a power data clustering method, a device and a storage medium based on power consumption behaviors, wherein the power data clustering method based on the power consumption behaviors comprises the following steps: extracting each evaluation index of each user according to historical electricity utilization data, and acquiring an evaluation index value and a weight coefficient corresponding to each evaluation index; obtaining an ideal solution of each evaluation index according to the evaluation index value corresponding to each user and the weight coefficient; acquiring the association degree between each user and the ideal solution, and acquiring the relative closeness of each user and the ideal solution according to the association degree; and obtaining the score of each user according to the relative closeness, and clustering the users according to the scores. By obtaining the association degree between each user and the ideal solution, the difference between each evaluation index and the ideal solution can be obtained, and the accuracy of cluster analysis is improved.
Description
Technical Field
The present application relates to the field of data calculation, and in particular, to a power consumption behavior-based power data clustering method, device, and storage medium.
Background
Energy is used as strategic resource, is the foundation and the life line of national economy and is also important material resource for the comprehensive development of society, and the construction of an intelligent management platform system is the core competitiveness of future deep informatization development of energy enterprises, and aims at realizing informatization of all links of production, transmission and consumption, and carrying out relevant linking, interconnection and intercommunication. The energy user is the terminal of energy industry chain, also is the key of energy enterprise's intelligent development upgrading, so, at the big datumization characteristic of electric power intelligent informatization construction stage that stands out day by day, it is essential core link to utilize data mining technique to dig latent analysis to user power consumption behavior characteristic, the analysis to user interaction power consumption behavior is the basis of formulating the demand response strategy that becomes more meticulous, the demand response is a user side response energy system side price signal or incentive mechanism to change inherent power consumption custom mode. However, at present, the traditional user electricity utilization behavior clustering is mostly limited to research and mining of load data, so that the clustering result is poor in practicability.
Disclosure of Invention
The embodiment of the application provides a power data clustering method, a device and a storage medium based on power consumption behaviors, and aims to solve the problem that the clustering of the power consumption behaviors of traditional users is mostly only limited to research and mining of load data, so that the clustering result is poor in practicability.
In order to achieve the above object, an aspect of the present application provides a power data clustering method based on power consumption behavior, the method including:
extracting each evaluation index of each user according to historical electricity utilization data, and acquiring an evaluation index value and a weight coefficient corresponding to each evaluation index;
obtaining an ideal solution of each evaluation index according to the evaluation index value corresponding to each user and the weight coefficient, wherein the ideal solution comprises a positive ideal solution and a negative ideal solution;
acquiring the association degree between each user and the ideal solution, and acquiring the relative closeness of each user and the ideal solution according to the association degree;
and obtaining the score of each user according to the relative closeness, and clustering the users according to the scores.
Optionally, the obtaining a degree of association between each user and the ideal solution, and the step of obtaining a relative closeness between each user and the ideal solution according to the degree of association includes:
acquiring a first association degree between the user and the positive ideal solution according to the evaluation index value corresponding to each user, and acquiring a second association degree between the user and the negative ideal solution according to the evaluation index value corresponding to each user;
and obtaining the relative closeness of each user to the positive ideal solution according to the first relevance degree and the second relevance degree corresponding to each user.
Optionally, the step of obtaining an ideal solution of each evaluation index according to the evaluation index value corresponding to each user and the weight coefficient includes:
constructing a weighting judgment matrix according to the evaluation index value and the weight coefficient corresponding to each user;
and acquiring the positive ideal solution and the negative ideal solution of each evaluation index according to the weighting judgment matrix.
Optionally, the step of obtaining a first degree of association between the user and the positive ideal solution according to the evaluation index value corresponding to each user, and obtaining a second degree of association between the user and the negative ideal solution according to the evaluation index value corresponding to each user includes:
acquiring a first Euclidean distance between each user and the positive ideal solution according to the difference value between each evaluation index value of each user and the positive ideal solution;
acquiring a second Euclidean distance between each user and the negative ideal solution according to the difference value between each evaluation index value of each user and the negative ideal solution;
and acquiring a first association degree between each user and the positive ideal solution according to the first Euclidean distance, and acquiring a second association degree between each user and the negative ideal solution according to the second Euclidean distance.
Optionally, the step of extracting the evaluation index of each user according to the historical electricity consumption data includes:
obtaining the extraction conditions of the evaluation indexes of at least two dimensions;
and extracting each evaluation index of each user from the historical electricity consumption data according to the extraction conditions.
Optionally, the dimensions include at least weather, economics, and policy.
Optionally, the step of obtaining the weighting factor corresponding to each evaluation index includes:
acquiring a subjective weight coefficient of each evaluation index;
obtaining an objective weight coefficient of each evaluation index;
and acquiring a weight coefficient corresponding to each evaluation index according to the subjective weight coefficient and the objective weight coefficient.
Optionally, the step of obtaining a subjective weight coefficient of each evaluation index includes:
classifying each evaluation index to obtain a hierarchical structure model of the evaluation index;
constructing a judgment matrix according to the hierarchical structure model, and carrying out consistency check on the judgment matrix;
when the consistency check passes, acquiring a subjective weight coefficient of each evaluation index according to the judgment matrix;
and correcting the judgment matrix when the consistency check fails.
Optionally, the step of obtaining an objective weight coefficient of each of the evaluation indexes includes:
acquiring an information entropy value of each evaluation index according to the evaluation index value corresponding to each user;
obtaining a difference coefficient of each evaluation index according to the information entropy;
and obtaining the objective weight coefficient of each evaluation index according to the difference coefficient.
Optionally, the step of performing consistency check on the determination matrix includes:
acquiring a preset characteristic value of the judgment matrix, and acquiring a consistency index according to the preset characteristic value;
obtaining a consistency ratio according to the consistency index;
and determining whether the consistency ratio is smaller than a preset value, wherein when the consistency ratio is smaller than the preset value, the consistency check is determined to be passed.
In addition, in order to achieve the above object, another aspect of the present application further provides a power data clustering device based on power consumption behaviors, the device includes a memory, a processor, and a power data clustering program stored in the memory and running on the processor, and the processor implements the steps of the power data clustering method based on power consumption behaviors as described above when executing the power data clustering program based on power consumption behaviors.
In addition, in order to achieve the above object, another aspect of the present application further provides a storage medium, on which a power consumption behavior-based power data clustering program is stored, and when the power consumption behavior-based power data clustering program is executed by a processor, the method of clustering power data based on power consumption behaviors as described above is implemented.
The embodiment of the application provides a power data clustering method based on power consumption behaviors, which comprises the steps of extracting each evaluation index of each user according to historical power consumption data, and obtaining evaluation index values and weight coefficients corresponding to the evaluation indexes; obtaining an ideal solution of each evaluation index according to the evaluation index value corresponding to each user and the weight coefficient; acquiring the association degree between each user and the ideal solution, and acquiring the relative closeness of each user and the ideal solution according to the association degree; and obtaining the score of each user according to the relative closeness, and clustering the users according to the scores. By obtaining the association degree between each user and the ideal solution, the difference between each evaluation index and the ideal solution can be obtained, and the accuracy of cluster analysis is improved.
Drawings
Fig. 1 is a schematic terminal structure diagram of a hardware operating environment according to an embodiment of the present application;
FIG. 2 is a schematic flow chart illustrating a first embodiment of a power consumption behavior-based power data clustering method according to the present application;
fig. 3 is a schematic flow chart illustrating the process of obtaining the weight coefficient corresponding to each evaluation index in the power data clustering method based on the power consumption behavior according to the present application;
FIG. 4 is a schematic diagram of a TOPSIS process defect;
FIG. 5 is a diagram illustrating the clustering result of the B2 dimension;
fig. 6 is a schematic diagram of the clustering result of the synthetic dimension.
The implementation, functional features and advantages of the objectives of the present application will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The main solution of the embodiment of the application is as follows: extracting each evaluation index of each user according to historical electricity utilization data, and acquiring an evaluation index value and a weight coefficient corresponding to each evaluation index; obtaining an ideal solution of each evaluation index according to the evaluation index value corresponding to each user and the weight coefficient, wherein the ideal solution comprises a positive ideal solution and a negative ideal solution; acquiring the association degree between each user and the ideal solution, and acquiring the relative closeness of each user and the ideal solution according to the association degree; and obtaining the score of each user according to the relative closeness, and clustering the users according to the scores.
Because the traditional user electricity utilization behavior clustering is mostly only limited to research and mining of load data, the clustering result is poor in practicability. Therefore, the application provides a power utilization behavior-based power data clustering method, which comprises the steps of extracting each evaluation index of each user according to historical power utilization data, and obtaining the evaluation index value and the weight coefficient corresponding to each evaluation index; obtaining an ideal solution of each evaluation index according to the evaluation index value corresponding to each user and the weight coefficient; acquiring the association degree between each user and the ideal solution, and acquiring the relative closeness of each user and the ideal solution according to the association degree; and obtaining the score of each user according to the relative closeness, and clustering the users according to the scores. By obtaining the association degree between each user and the ideal solution, the difference between each evaluation index and the ideal solution can be obtained, and the accuracy of cluster analysis is improved.
As shown in fig. 1, fig. 1 is a schematic terminal structure diagram of a hardware operating environment according to an embodiment of the present application.
As shown in fig. 1, the terminal may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, a communication bus 1002. Wherein a communication bus 1002 is used to enable connective communication between these components. The user interface 1003 may include a Display screen (Display), an input unit such as a Keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface, a wireless interface. The network interface 1004 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface). The memory 1005 may be a high-speed RAM memory or a non-volatile memory (e.g., a magnetic disk memory). The memory 1005 may alternatively be a storage device separate from the processor 1001.
Those skilled in the art will appreciate that the terminal structure shown in fig. 1 does not constitute a limitation of the terminal device and may include more or fewer components than those shown, or some components may be combined, or a different arrangement of components.
As shown in fig. 1, a memory 1005, which is a kind of computer-readable storage medium, may include therein an operating system, a network communication module, a user interface module, and a power usage behavior-based power data clustering program.
In the terminal shown in fig. 1, the network interface 1004 is mainly used for data communication with the background server; the user interface 1003 is mainly used for data communication with a client (user side); the processor 1001 may be configured to invoke a power usage behavior-based power data clustering routine in the memory 1005 and perform the following operations:
extracting each evaluation index of each user according to historical electricity utilization data, and acquiring an evaluation index value and a weight coefficient corresponding to each evaluation index;
obtaining an ideal solution of each evaluation index according to the evaluation index value corresponding to each user and the weight coefficient, wherein the ideal solution comprises a positive ideal solution and a negative ideal solution;
acquiring the association degree between each user and the ideal solution, and acquiring the relative closeness of each user and the ideal solution according to the association degree;
and obtaining the score of each user according to the relative closeness, and clustering the users according to the scores.
Referring to fig. 2, fig. 2 is a schematic flowchart of a first embodiment of the power consumption behavior-based power data clustering method according to the present application.
The embodiments of the present application provide a power data clustering method based on power consumption behavior, and it should be noted that, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in an order different from that here.
The power data clustering method based on the power utilization behaviors comprises the following steps:
step S10, extracting each evaluation index of each user according to historical electricity utilization data, and acquiring the evaluation index value and the weight coefficient corresponding to each evaluation index;
the method comprises the steps of collecting historical electricity utilization data of a user, such as electricity utilization time, electricity consumption, electricity quantity price and other data, analyzing the historical electricity utilization data, and extracting and defining electricity utilization behavior characteristic key sub-indexes, namely evaluation indexes, from multiple dimensions, wherein the multiple dimensions refer to dimensions related to electricity utilization behaviors of the user, such as weather, economy, policies and the like. In one embodiment, a plurality of evaluation indexes in the comprehensive information of the actual power consumption and the subjective and objective conditional willingness of the user are extracted from the historical power consumption data through a statistical method, and as the meaning of each evaluation index is different from the application scene, the corresponding extraction conditions are also different, for example: the electricity consumption actual data index is to extract a peak value from the actual data, and the green index sets corresponding formula quantization for the government policy index. Therefore, it is necessary to obtain extraction conditions of the evaluation index of each dimension, such as an extraction formula, extraction peak information, and the like, and then extract each evaluation index of each user, such as a reliability index, an energy saving and environmental protection index, an economic benefit index, and the like, from the historical electricity data according to the extraction conditions. And simultaneously, defining and measuring and calculating methods of all evaluation indexes are provided, so that a multi-dimensional index evaluation analysis system, namely an index system, of the user power consumption behavior correlation characteristic factors is established, wherein the power consumption behavior correlation factors influence the user power consumption behavior correlation factors. After the index system is established, the scoring standard of each evaluation index of the electricity user needs to be determined, for example, the evaluation grade and the standard are formulated, the unified evaluation grade or the score range of each evaluation index is formulated firstly, and then the standard of each grade of each evaluation index is formulated. In this way, the evaluation index value corresponding to each evaluation index is acquired by determining the scoring standard of each evaluation index of the electricity user.
Because different dimensions, such as time, length, quality and the like, exist among different evaluation indexes, so that the different evaluation indexes are not comparable, the evaluation indexes are subjected to non-dimensionalization (normalization of data) firstly, the weight coefficient corresponding to each evaluation index is obtained after the influence of the dimensions is eliminated, and the weight coefficient refers to the importance degree of each evaluation index in an index system and is endowed with a corresponding value. In an embodiment, in the aspect of an evaluation method, an entropy weight method is mostly adopted in the existing evaluation method, the entropy weight method is an objective weighting method, a phenomenon of partial evaluation index weight unbalance may occur, meanwhile, subjective weighting cannot be performed according to specific conditions of different power utilization users, and certain limitations exist. Therefore, in this embodiment, subjective assignment is performed on each evaluation index in the index system by an Analytic Hierarchy Process (AHP) to obtain a subjective weight of each evaluation index; meanwhile, each evaluation index in the index system is objectively assigned through an entropy weight method, and the objective weight of each evaluation index is obtained. Further, the comprehensive weight of each evaluation index is calculated through the subjective weight and the objective weight of each evaluation index, wherein in order to overcome the problem that the subjective factor of the comprehensive weight obtained by the commonly-used linear addition integration weighting is too strong and correctly process the preference factor of the subjective and objective weights, the embodiment adopts a multiplication integration weighting method to calculate the comprehensive weight, and the corresponding calculation formula is as follows:
wherein j is an evaluation index, ajSubjective weight as the j-th evaluation index, bjIs the objective weight of the j-th evaluation index, wjAnd the j-th evaluation index is the comprehensive weight of the j-th evaluation index.
The Analytic Hierarchy Process (AHP) is a systematic method which takes a complex multi-target decision problem as a system, decomposes a target into a plurality of targets or criteria and further decomposes the targets into a plurality of layers of multi-index (or criteria and constraints), and calculates the single-layer ordering (weight) and the total ordering by a qualitative index fuzzy quantization method to be taken as the target (multi-index) and multi-scheme optimization decision. The entropy weight method is an objective weighting method, and the basic idea is to determine objective weight according to the index variability, and the principle is as follows: the smaller the variation degree of the index is, the less the amount of information is reflected, and the lower the corresponding weight value is.
Step S20, obtaining an ideal solution of each evaluation index according to the evaluation index value corresponding to each user and the weight coefficient, wherein the ideal solution comprises a positive ideal solution and a negative ideal solution;
it should be noted that, in the GRA-TOPSIS comprehensive evaluation model adopted in the present application, n evaluation indexes are regarded as n coordinate axes, so that an n-dimensional space can be constructed, each object to be evaluated corresponds to one coordinate point in the n-dimensional space according to data of each evaluation index, and an optimal value (a positive ideal solution, corresponding to an optimal coordinate point) and a worst value (a negative ideal solution, corresponding to a worst coordinate point) of each evaluation index are selected from all objects to be evaluated according to each evaluation index. The ideal solutions in this embodiment include a positive ideal solution and a negative ideal solution, where obtaining the positive ideal solution and the negative ideal solution corresponding to each evaluation index according to the evaluation index value corresponding to each user and the weight coefficient includes the following steps:
step 1: constructing a weighted decision moment rijAnd constructing a weighted judgment matrix based on the electricity utilization user sample evaluation matrix data of the dimensionless standard and the subjective and objective weighting model facing the electricity utilization behavior association characteristic factor multi-dimensional index. Namely:
rij＝x’ij×wj,i＝1,2,...,m；j＝1,2,...,n
in formula (II), x'ijAnd j is an evaluation index and represents a standard value obtained by carrying out non-dimensionalization on the original electricity user sample evaluation matrix data.
Step 2: determining a positive ideal solution for a user
Sum negative ideal solution
By a dimensionless weighting matrix rijObtaining an extreme value type index of the user evaluation matrix, namely:
wherein, for
In other words, if the jth evaluation index is a forward evaluation index, the larger the positive ideal solution value, the better; if the jth evaluation index is a negative evaluation index, the smaller the positive ideal solution value is, the better the positive ideal solution value is. For the
For example, if the jth evaluation index is a positive evaluation index, the larger the negative ideal solution value is, the better it is; if the jth evaluation index is a negative evaluation index, the smaller the negative ideal solution value is, the better the negative ideal solution value is.
Step S30, obtaining the association degree between each user and the ideal solution, and obtaining the relative closeness degree between each user and the ideal solution according to the association degree;
the method is characterized in that a conventional TOPSIS method is a multi-target decision ordering method approaching to an ideal solution, the basic ordering idea is to draw up a positive ideal evaluation object and a negative ideal evaluation object on the basis of a weighted dimensionless normalized evaluation matrix, determine Euclidean distances between each evaluation sample and the drawn positive ideal evaluation object and negative ideal evaluation object respectively, and finally order the evaluation objects according to the relative degree of progress of each evaluation object.
As can be seen from the basic sorting thought of the TOPSIS method, there is a disadvantage that the result of sorting the evaluation objects according to the euclidean distance does not completely reflect the characteristic difference of each object, that is, a scheme closer to the positive ideal scheme may be closer to the euclidean distance of the negative ideal scheme, and only the relative closeness of each evaluation object can be reflected, as shown in fig. 4, which is a point shown in fig. 4
And point
Respectively representing a negative ideal evaluation object and a positive ideal evaluation object to be selected, and a, b, c and d respectively representing other sample objects to be sorted. a. b evaluating the object to
The relative closeness is consistent, and the evaluation indexes inside the samples a and b may be far apart. Similarly, samples c and d could not be distinguished by TOPSIS.
Based on the above thought, in the embodiment, aiming at the defect that the TOPSIS method can reflect the similarity between each user and the positive and negative ideal objects by calculating the relative closeness of each object in the multi-object power user evaluation problem, but cannot reflect the change of each evaluation characteristic index inside the power user and the difference between the positive and negative ideal objects, the gray correlation degree is introduced to analyze the difference between the change situation of each evaluation factor inside the power user and the positive and negative ideal schemes, and the power consumption behavior correlation characteristic factor oriented multidimensional GRA-TOPSIS comprehensive evaluation model is constructed. The method comprises the following steps of obtaining the association degree between each user and an ideal solution through a GRA-TOPSIS comprehensive evaluation model, and obtaining the relative closeness degree between each user and the ideal solution according to the association degree, wherein the method comprises the following steps:
step 1: calculating the Euclidean form of the user to the positive and negative ideal value solution object
Step 2: calculating the relevance matrix xi of the user and the solution sample of the positive and negative ideal values+And xi-. Calculating the correlation coefficient xi between the user and the positive and negative ideal value solution samplesijThen according to the correlation coefficient xiijCalculating the matrix xi of the degree of association+And xi-Wherein the correlation coefficient ξijPositive correlation coefficient matrix xi+And negative correlation coefficient matrix xi-As follows:
in the formula, xiijIndicating the degree of association, Δ, between i-household and the reference object with respect to the j-th indexi(j)＝|r0j-rijL, represents r0And riThe absolute difference at the k-th index, rho, is a resolution coefficient, and the main purpose is to reduce the influence of distortion due to too large absolute difference, wherein the value range of rho is usually 0-1, and rho is 0.5.
And step 3: calculating the closeness degree of the user sample to the positive and negative ideal solutions
Combined dimensionless Euclidean distance diDegree of association with gray σiObtaining:
in the formula, λ1+λ 21, the coefficient represents the preference of the decision maker, λ1And λ2Are all 0.5;
coefficient matrix xi taken from negative correlation-，
From the positive correlation coefficient matrix xi+。
And 4, step 4: calculating the relative closeness D of the user object to the positive ideal solutioni(x)：
Wherein S includes the proximity of the user to the ideal solution
And the proximity of the user to the negative ideal solution
According to relative closeness Di(x) Size, ranking the users to be evaluated, and relative closeness Di(x) The larger the evaluation, the closer the user is to the positive ideal solution, and the more the ranking is forward; conversely, the more backward the ranking.
And step S40, obtaining the score of each user according to the relative closeness, and clustering the users according to the scores.
The cluster analysis is an analysis process of grouping a set of physical or abstract objects into a plurality of classes composed of similar objects, and is a multivariate statistical analysis method for quantitatively analyzing a plurality of samples.
And ordering the users according to the values of the relative closeness, grading the users based on the ordering result, and clustering the users according to the grades, for example, classifying the users with similar grades into a class, wherein a K-means clustering algorithm, a hierarchical clustering algorithm, an SOM clustering algorithm, an FCM clustering algorithm and the like can be adopted during clustering. In one embodiment, the dimensions are from four different dimensions: the method comprises the steps of carrying out evaluation, sequencing, clustering and analysis on 10 non-resident user real electricity consumption related data serving as research objects by using an electricity load complex weather related factor index (B1), an actual electricity load characteristic factor index (B2), an energy-saving environment-friendly green factor index (B3) and a user interaction adjustment potential factor (B4) to obtain user clustering results with different dimensions. Referring to fig. 5, fig. 5 shows the clustering result of the B2 dimension, and it can be seen from fig. 5 that the users can be classified into three types in the B2 dimension. In the embodiment, multivariate information resources are considered, intelligent terminal data and power consumption behavior subjective and objective power consumption associated information are integrated, power consumption behavior associated characteristic factor indexes are considered in multiple angles, and in the practical application process, different dimensionality evaluation systems can be established in a key mode according to the attention research direction, and then value information is mined.
Referring to fig. 6, as can be seen from the content shown in fig. 6, based on the GRA-TOPSIS multi-dimensional index comprehensive evaluation clustering model, considering four dimensional index characteristics at the same time, users can be classified into three categories, the user type I comprehensive characteristic index is higher than the user type II and the user type III, wherein the user 23 comprehensive characteristic index is the highest, and the user 25 is the lowest. In actual needs, multi-scale standard clustering can be performed on users according to factors such as service directions.
The GRA-TOPSIS evaluation ranking refined clustering model based on the multidimensional index of the power utilization behavior of the user is connected with the information of the multiple power utilization characteristic correlation factors, so that the clustering of different dimensions of power users is effectively realized, and references can be provided for energy suppliers to efficiently customize services in different directions. Meanwhile, the clustering result can be used for analyzing the specific characteristics of different users in the same type of users with different dimensions, for example, taking the dimension clustering result oriented to the interactive adjustment potential factor as an example, when a user type I user with the largest relative adjustment potential is selected as a service object, if the similarity of the actual power load characteristics of different users in the type I user needs to be further researched, the rough information of the power load characteristics of the users can be obtained according to the dimension index coefficient B2 in fig. 5. In practical situations, under the background of giving priority to a certain dimension of electricity utilization correlation characteristic indexes, an energy company comprehensively considers other dimension factors and performs value-added quality services on different power consumers.
According to the embodiment, the power utilization information such as the adjustment potential and the load power of the power consumer is mined based on the power load characteristic analysis by applying a data mining technology, and meanwhile, the power utilization information is not limited to the power load data analysis of the user, and the evaluation method is applied to carry out multi-dimensional fine clustering research on the power utilization behavior of the power consumer by integrating the power utilization behavior associated characteristic factor indexes of the user. Thus, the present application has the following advantages:
the method is not limited to load data research and mining, and the user power consumption behaviors are clustered from multiple angles by analyzing the multi-dimensional power consumption behavior association factors of the user;
a comprehensive subjective and objective weighting method based on a hierarchical analysis subjective weighting method and an information entropy objective weighting method is adopted to weight the multidimensional indexes by using the associated characteristic factors of the electricity consumption behaviors of the users, so that the single weighting sheet sidedness is improved;
based on the comprehensive weighting model, the shortcomings of the traditional TOPSIS method are improved by utilizing the grey correlation degree, and a GRA-TOPSIS comprehensive evaluation refined clustering model facing to the multidimensional index of the user electricity consumption behavior correlation characteristic factors is provided, so that the accuracy of clustering analysis is improved.
Further, referring to fig. 3, a second embodiment of the power data clustering method based on power consumption behavior is provided.
The second embodiment of the power consumption behavior-based power data clustering method is different from the first embodiment in that the step of obtaining the weight coefficient corresponding to each evaluation index comprises the following steps:
step S11, acquiring the subjective weight coefficient of each evaluation index;
based on the subjective weighting of the hierarchical analysis, the multi-dimensional indexes of the user electricity consumption behavior associated characteristic factors are classified by using the hierarchical analysis method to form a hierarchical structure model, wherein if the first layer is a first-level evaluation index; the second layer is a secondary evaluation index, and so on. Obtaining the evaluation index of each layer in the hierarchical structure model, constructing a judgment matrix A in a pairwise comparison mode, setting n evaluation indexes, determining the relative importance of each element in the same layer through the judgment matrix A, and further obtaining the corresponding subjective weight coefficient. Wherein, the judgment matrix A is:
in the formula, A is a judgment matrix corresponding to the index of the previous level; a isjIs the relative importance between sub-primary indexes, i.e. the relative weight, ajAnd the subjective weight coefficient of the j-th evaluation index. And calculating according to the judgment matrix to obtain the weight coefficient of each level aiming at the previous level, wherein an assignment scale principle is formulated by adopting the following table 1:
TABLE 1 assignment Scale principles
| Dimensionwij | Means of |
| 1 | CiAnd CjHas the same effect |
| 3 | CiRatio CjIs slightly more strongly influenced |
| 5 | CiRatio CjHas strong influence on |
| 7 | CiRatio CjIs obviously strong |
| 9 | CiRatio CjIs absolutely strong |
| 2，4，6，8 | CiAnd CjBetween the above two |
| 1，1/2，…，1/9 | CiAnd CjHas an influence of wijReciprocal number of |
In order to ensure the reasonability of the judgment matrix, the consistency of the judgment matrix needs to be checked, and when the judgment matrix meets the consistency requirement, the subjective weight coefficient a of each index can be determinedj. The consistency check of the judgment matrix comprises the following steps:
step 1: calculating consistency check index CI, vmaxIs the maximum feature of the decision matrixAnd p is the number of the sub items of the index layer, wherein the calculation formula of CI is as follows:
step 2: inquiring the average random consistency index RI corresponding to n;
and step 3: calculating a consistency ratio CR, and if CR is less than 0.1, judging that the matrix is consistent with consistency; if CR >0.1, the decision matrix needs to be modified so that CR <0.1, wherein the formula for calculating the consistency ratio CR is as follows:
and 4, step 4: and when the consistency requirement is met, obtaining the subjective weight coefficient of each evaluation index after normalizing the feature vector.
Step S12, obtaining an objective weight coefficient of each evaluation index;
the entropy-based objective weighting method is a weighting method for determining the weight of the evaluation index by applying an entropy weighting method according to the information content richness and the difference degree contained in each multi-dimensional evaluation index of the user electricity consumption behavior associated characteristic factors. The entropy weight method refers to the entropy value in the information theory to represent the disorder degree of the information in different evaluation indexes, so that the quantity of the information contained in a certain evaluation index and the difference degree are measured, and the action of the evaluation index on the target decision is determined.
The user electricity consumption behavior correlation characteristic factor index matrix is set as follows:
R＝{xij|i＝1,2,...,m；j＝1,2,...,n}
in the formula, m is the number of users using electricity, n is the number of evaluation index items, and xijThe objective weighting steps based on the information entropy for the jth evaluation index value of the ith user are as follows:
step 1: calculating an information entropy value I of an evaluation index jj：
In the formula, yijThe characteristic proportion of the j-th evaluation index of the ith user.
Step 2: calculating the difference coefficient H of the evaluation index jj：
Hj＝1-Ij,j＝1,2,...n
Wherein HjReflecting the degree of discrimination of the evaluation index j to different users, HjThe larger the evaluation index is, the larger the discrimination of the evaluation index to different evaluation objects is, the better the effect of adopting the evaluation index is.
And step 3: calculating the weight coefficient b of the evaluation index jjThe ratio of the evaluation index j to the sum of all the n difference coefficients is used as the weight coefficient of the evaluation index j, and the calculation formula is as follows:
and step S13, obtaining a weight coefficient corresponding to each evaluation index according to the subjective weight coefficient and the objective weight coefficient.
The comprehensive subjective and objective weighting method based on the hierarchical analysis subjective weighting method and the information entropy objective weighting method is adopted to weight the multidimensional index by the user electricity consumption behavior associated characteristic factor, so that the subjective experience research and judgment of a decision maker and the objective and real data samples are considered, and the scientific and reasonable weighting is ensured. In order to overcome the problem of too strong subjective factors of the comprehensive weight obtained by the conventional linear addition integrated weighting and correctly process the preference factors of the subjective and objective weights, the comprehensive weight is calculated by adopting a multiplication integrated weighting method, and the corresponding calculation formula is as follows:
wherein j is an evaluation index, ajSubjective weight as the j-th evaluation index, bjIs the objective weight of the j-th evaluation index, wjAnd the j-th evaluation index is the comprehensive weight of the j-th evaluation index.
In the embodiment, each evaluation index is assigned by adopting a comprehensive subjective and objective weighting method based on a hierarchical analysis subjective weighting method and an information entropy objective weighting method, so that the scientific rationality of weighting is ensured.
In addition, the application also provides a power data clustering device based on the power utilization behaviors, the device comprises a memory, a processor and a power data clustering program which is stored on the memory and runs on the processor based on the power utilization behaviors, and the device extracts each evaluation index of each user according to historical power utilization data to obtain an evaluation index value and a weight coefficient corresponding to each evaluation index; obtaining an ideal solution of each evaluation index according to the evaluation index value corresponding to each user and the weight coefficient; acquiring the association degree between each user and the ideal solution, and acquiring the relative closeness of each user and the ideal solution according to the association degree; and obtaining the score of each user according to the relative closeness, and clustering the users according to the scores. By obtaining the association degree between each user and the ideal solution, the difference between each evaluation index and the ideal solution can be obtained, and the accuracy of cluster analysis is improved.
In addition, the present application also provides a storage medium, on which a power consumption behavior-based power data clustering method program is stored, and when being executed by a processor, the power consumption behavior-based power data clustering method program implements the steps of the power consumption behavior-based power data clustering method described above.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It should be noted that in the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The application can be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by one and the same item of hardware. The usage of the words first, second and third, etcetera do not indicate any ordering. These words may be interpreted as names.
While alternative embodiments of the present application have been described, additional variations and modifications of these embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. It is therefore intended that the following appended claims be interpreted as including alternative embodiments and all such alterations and modifications as fall within the scope of the application.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.
Claims (12)
1. A power data clustering method based on power consumption behaviors is characterized by comprising the following steps:
extracting each evaluation index of each user according to historical electricity utilization data, and acquiring an evaluation index value and a weight coefficient corresponding to each evaluation index;
obtaining an ideal solution of each evaluation index according to the evaluation index value corresponding to each user and the weight coefficient, wherein the ideal solution comprises a positive ideal solution and a negative ideal solution;
acquiring the association degree between each user and the ideal solution, and acquiring the relative closeness of each user and the ideal solution according to the association degree;
and obtaining the score of each user according to the relative closeness, and clustering the users according to the scores.
2. The method for clustering power data based on electricity consumption behaviors according to claim 1, wherein the step of obtaining the association degree between each user and the ideal solution comprises the steps of:
acquiring a first association degree between the user and the positive ideal solution according to the evaluation index value corresponding to each user, and acquiring a second association degree between the user and the negative ideal solution according to the evaluation index value corresponding to each user;
and obtaining the relative closeness of each user to the positive ideal solution according to the first relevance degree and the second relevance degree corresponding to each user.
3. The power consumption behavior-based power data clustering method according to claim 1, wherein the step of obtaining an ideal solution for each of the evaluation indexes according to the evaluation index value corresponding to each user and the weight coefficient comprises:
constructing a weighting judgment matrix according to the evaluation index value and the weight coefficient corresponding to each user;
and acquiring the positive ideal solution and the negative ideal solution of each evaluation index according to the weighting judgment matrix.
4. The power consumption behavior-based power data clustering method according to claim 2, wherein the step of obtaining a first degree of association between the user and the positive ideal solution according to the evaluation index value corresponding to each user, and obtaining a second degree of association between the user and the negative ideal solution according to the evaluation index value corresponding to each user comprises:
acquiring a first Euclidean distance between each user and the positive ideal solution according to the difference value between each evaluation index value of each user and the positive ideal solution;
acquiring a second Euclidean distance between each user and the negative ideal solution according to the difference value between each evaluation index value of each user and the negative ideal solution;
and acquiring a first association degree between each user and the positive ideal solution according to the first Euclidean distance, and acquiring a second association degree between each user and the negative ideal solution according to the second Euclidean distance.
5. The power consumption behavior-based power data clustering method according to claim 1, wherein the step of extracting the individual evaluation index of each user from the historical power consumption data comprises:
obtaining the extraction conditions of the evaluation indexes of at least two dimensions;
and extracting each evaluation index of each user from the historical electricity consumption data according to the extraction conditions.
6. The method according to claim 5, wherein the dimensions include at least weather, economy, and policy.
7. The power consumption behavior-based power data clustering method according to claim 1, wherein the step of obtaining the weight coefficient corresponding to each evaluation index comprises:
acquiring a subjective weight coefficient of each evaluation index;
obtaining an objective weight coefficient of each evaluation index;
and acquiring a weight coefficient corresponding to each evaluation index according to the subjective weight coefficient and the objective weight coefficient.
8. The method for clustering power data based on electricity consumption behavior according to claim 7, wherein the step of obtaining a subjective weighting factor for each of the evaluation indexes comprises:
classifying each evaluation index to obtain a hierarchical structure model of the evaluation index;
constructing a judgment matrix according to the hierarchical structure model, and carrying out consistency check on the judgment matrix;
when the consistency check passes, acquiring a subjective weight coefficient of each evaluation index according to the judgment matrix;
and correcting the judgment matrix when the consistency check fails.
9. The power consumption behavior-based power data clustering method according to claim 7, wherein the step of obtaining the objective weight coefficient for each of the evaluation indexes comprises:
acquiring an information entropy value of each evaluation index according to the evaluation index value corresponding to each user;
obtaining a difference coefficient of each evaluation index according to the information entropy;
and obtaining the objective weight coefficient of each evaluation index according to the difference coefficient.
10. The power consumption behavior-based power data clustering method according to claim 8, wherein the step of performing consistency check on the judgment matrix comprises:
acquiring a preset characteristic value of the judgment matrix, and acquiring a consistency index according to the preset characteristic value;
obtaining a consistency ratio according to the consistency index;
and determining whether the consistency ratio is smaller than a preset value, wherein when the consistency ratio is smaller than the preset value, the consistency check is determined to be passed.
11. An electricity usage behavior based power data clustering apparatus, characterized in that the apparatus comprises a memory, a processor and a power data clustering program stored on the memory and running on the processor, the processor implementing the steps of the method according to any one of claims 1 to 10 when executing the power data clustering program.
12. A storage medium, characterized in that the storage medium has stored thereon a power usage behavior based power data clustering program, which when executed by a processor implements the steps of the method according to any one of claims 1 to 10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110804808.5A CN113591947A (en) | 2021-07-15 | 2021-07-15 | Power data clustering method and device based on power consumption behaviors and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110804808.5A CN113591947A (en) | 2021-07-15 | 2021-07-15 | Power data clustering method and device based on power consumption behaviors and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113591947A true CN113591947A (en) | 2021-11-02 |
Family
ID=78248319
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110804808.5A Pending CN113591947A (en) | 2021-07-15 | 2021-07-15 | Power data clustering method and device based on power consumption behaviors and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113591947A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115829418A (en) * | 2023-02-07 | 2023-03-21 | 国网江苏省电力有限公司营销服务中心 | Power consumer load characteristic portrait construction method and system suitable for load management |
| CN116188622A (en) * | 2022-12-08 | 2023-05-30 | 北京国电通网络技术有限公司 | Power index information display method and device, electronic equipment and computer medium |
-
2021
- 2021-07-15 CN CN202110804808.5A patent/CN113591947A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116188622A (en) * | 2022-12-08 | 2023-05-30 | 北京国电通网络技术有限公司 | Power index information display method and device, electronic equipment and computer medium |
| CN116188622B (en) * | 2022-12-08 | 2023-09-12 | 北京国电通网络技术有限公司 | Power index information display method and device, electronic equipment and computer medium |
| CN115829418A (en) * | 2023-02-07 | 2023-03-21 | 国网江苏省电力有限公司营销服务中心 | Power consumer load characteristic portrait construction method and system suitable for load management |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110866819A (en) | Automatic credit scoring card generation method based on meta-learning | |
| CN111324642A (en) | Model algorithm type selection and evaluation method for power grid big data analysis | |
| CN110866782B (en) | Customer classification method and system and electronic equipment | |
| CN113591947A (en) | Power data clustering method and device based on power consumption behaviors and storage medium | |
| CN114880486A (en) | Industry chain identification method and system based on NLP and knowledge graph | |
| CN109726918A (en) | The personal credit for fighting network and semi-supervised learning based on production determines method | |
| CN109783633A (en) | Data analysis service procedural model recommended method | |
| CN112989621A (en) | Model performance evaluation method, device, equipment and storage medium | |
| CN112598405B (en) | Business project data management method and system based on big data | |
| CN111612583B (en) | Personalized shopping guide system based on clustering | |
| KR101259417B1 (en) | Hybrid type method and system for extracting a emerging technologies using collective intelligence | |
| TW201243627A (en) | Multi-label text categorization based on fuzzy similarity and k nearest neighbors | |
| CN112508363A (en) | Deep learning-based power information system state analysis method and device | |
| CN115829683A (en) | Power integration commodity recommendation method and system based on inverse reward learning optimization | |
| CN114372835B (en) | Comprehensive energy service potential customer identification method, system and computer equipment | |
| CN114004530B (en) | Enterprise electric power credit modeling method and system based on ordering support vector machine | |
| CN115271442A (en) | Modeling method and system for evaluating enterprise growth based on natural language | |
| CN113705920A (en) | Generation method of water data sample set for thermal power plant and terminal equipment | |
| Akyol | Clustering hotels and analyzing the importance of their features by machine learning techniques | |
| CN111127184A (en) | Distributed combined credit evaluation method | |
| CN112418987A (en) | Method and system for rating credit of transportation unit, electronic device and storage medium | |
| CN112288571A (en) | Personal credit risk assessment method based on rapid construction of neighborhood coverage | |
| CN114281994B (en) | Text clustering integration method and system based on three-layer weighting model | |
| CN117151870B (en) | Portrait behavior analysis method and system based on guest group | |
| CN117035837B (en) | Method for predicting electricity purchasing demand of power consumer and customizing retail contract |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information |
Address after: 230088 101, 201, office building of Hefei Haoyang Power Technology Co., Ltd., No.23 Tianzhi Road, high tech Zone, Hefei City, Anhui Province Applicant after: Sunshine Zhiwei Technology Co.,Ltd. Address before: 230088 101, 201, office building of Hefei Haoyang Power Technology Co., Ltd., No.23 Tianzhi Road, high tech Zone, Hefei City, Anhui Province Applicant before: Sunshine Zhiwei Technology Co.,Ltd. Address after: 230088 101, 201, office building of Hefei Haoyang Power Technology Co., Ltd., No.23 Tianzhi Road, high tech Zone, Hefei City, Anhui Province Applicant after: Sunshine Zhiwei Technology Co.,Ltd. Address before: 230088 101, 201, office building of Hefei Haoyang Power Technology Co., Ltd., No.23 Tianzhi Road, high tech Zone, Hefei City, Anhui Province Applicant before: Hefei sunshine Zhiwei Technology Co.,Ltd. |
|
| CB02 | Change of applicant information |