Power grid sensitive user identification method and device based on multi-dimensional data perception and medium

Abstract

The invention discloses a power grid sensitive user identification method and device based on multi-dimensional data perception and a medium, the method avoids one-sidedness of power grid sensitive user identification based on single data through multi-dimensional data fusion of basic attributes, power consumption behaviors, complaint information, service information and the like, improves the reliability of power grid sensitive user identification, and improves the power grid sensitive user identification efficiency. Automatic user identification is achieved through the sensitive user classification model, manual participation in the identification process is not needed, the power grid sensitive user identification efficiency is improved, sensitive user identification can be conducted through multi-dimensional data under the condition that complaint information does not exist, the hysteresis quality that part of sensitive user identification can only be completed after complaints are received is avoided, and the user identification efficiency is improved. Therefore, the complaint rate and the user loss are effectively reduced, and the refinement degree of power grid user management is improved.

Description

Technical Field

[0001] This invention relates to the field of power grid user identification, and in particular to a method, device and medium for identifying power grid sensitive users based on multi-dimensional data perception. Background Technology

[0002] With the deepening of smart grid construction, the types of power grid users are becoming increasingly diversified, and different user types have significantly different sensitivities to power supply reliability and stability. Traditional methods for identifying sensitive power grid users have obvious limitations: First, they rely on single-dimensional data, such as classifying sensitive users solely based on user type, leading to poor reliability in identifying sensitive power grid users. Second, they are primarily based on manual judgment, which is inefficient, highly subjective, and difficult to adapt to the needs of accurate identification of massive numbers of users.

[0003] Therefore, improving the reliability and efficiency of identifying sensitive users in the power grid has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a method, device, and medium for identifying power grid sensitive users based on multi-dimensional data perception, in order to solve the problems of poor reliability and efficiency in identifying power grid sensitive users.

[0005] A method for identifying power grid sensitive users based on multi-dimensional data perception, the method comprising:

[0006] Obtain comprehensive attribute information of the user to be identified, wherein the comprehensive attribute information includes at least basic attribute information, electricity consumption behavior information, historical complaint information and historical service information;

[0007] The comprehensive attribute information is input into the trained sensitive user classification model to obtain the sensitive user classification result corresponding to the user to be identified;

[0008] When the sensitive user classification result corresponding to the user to be identified meets the preset conditions, the user to be identified is determined to be a power grid sensitive user.

[0009] A method and apparatus for identifying power grid sensitive users based on multi-dimensional data perception, the method and apparatus comprising:

[0010] A multi-dimensional information acquisition module is used to acquire comprehensive attribute information of the user to be identified, wherein the comprehensive attribute information includes at least basic attribute information, electricity consumption behavior information, historical complaint information and historical service information;

[0011] The model inference module is used to input the comprehensive attribute information into the trained sensitive user classification model to obtain the sensitive user classification result corresponding to the user to be identified;

[0012] The sensitive user identification module is used to determine that the user to be identified is a power grid sensitive user when the sensitive user classification result corresponding to the user to be identified meets the preset conditions.

[0013] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the above-described method for identifying power grid sensitive users based on multidimensional data perception.

[0014] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for identifying power grid-sensitive users based on multidimensional data perception.

[0015] The aforementioned method, equipment, and medium for identifying sensitive users of the power grid based on multi-dimensional data perception, through the fusion of multi-dimensional data such as basic attributes, electricity consumption behavior, complaint information, and service information, avoids the one-sidedness of identifying sensitive users of the power grid based on single data, improves the reliability of identifying sensitive users of the power grid, and achieves automated user identification through a sensitive user classification model without the need for manual intervention in the identification process, thereby improving the efficiency of identifying sensitive users of the power grid. Moreover, it can identify sensitive users through multi-dimensional data even without complaint information, avoiding the lag in identifying some sensitive users that can only be completed after a complaint is received, thereby effectively reducing the complaint rate and user losses, and improving the precision of power grid user management. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of an application environment for a power grid sensitive user identification method based on multi-dimensional data perception in one embodiment of the present invention;

[0018] Figure 2 This is a flowchart of a power grid sensitive user identification method based on multi-dimensional data perception in one embodiment of the present invention;

[0019] Figure 3 This is a schematic diagram of a device for identifying power grid sensitive users based on multi-dimensional data perception in one embodiment of the present invention;

[0020] Figure 4 This is a schematic diagram of a computer device according to an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] The power grid sensitive user identification method based on multi-dimensional data perception provided in this invention can be applied to, for example... Figure 1 The application environment is shown. Specifically, this method for identifying power grid-sensitive users based on multi-dimensional data perception is applied in a user data management system within a power grid scenario. This user data management system includes, for example, […]. Figure 1 The diagram shows a client and a server, with the client communicating with the server over a network. The client, also known as the user terminal, refers to the program that provides local services to the client, corresponding to the server. The client can be installed on, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server can be implemented using a standalone server or a server cluster consisting of multiple servers.

[0023] In one embodiment, such as Figure 2 As shown, a method for identifying power grid-sensitive users based on multi-dimensional data perception is provided, and this method is applied to... Figure 1 Taking the server in the example, the following steps are included:

[0024] Step S201: Obtain comprehensive attribute information of the user to be identified, wherein the comprehensive attribute information includes at least basic attribute information, electricity consumption behavior information, historical complaint information, and historical service information.

[0025] Among them, comprehensive attribute information can refer to the collection of core data used to describe user characteristics. Basic attribute information may include user identity information, user type information, user geographical location information, user electricity consumption level information, etc. User type information may include hospitals, schools, enterprises, etc.

[0026] Electricity consumption behavior information may include average power supply, load fluctuation coefficient, peak-valley electricity consumption difference, average electricity consumption duration, etc. Historical complaint information may include the number of complaints, complaint type, complaint time, etc. Historical service information may include the number of repair requests, complaint resolution satisfaction, service interaction frequency, etc.

[0027] Specifically, the comprehensive attribute information can be represented by a vector with 1 row and Q columns. Each column in the vector corresponds to information of one dimension. For example, the first column of the vector can represent user identity information, the second column of the vector can represent user type information, etc. The information of each dimension is digitized through digital encoding. For example, the digital encoding corresponding to user type information is 0, 1, 2, where 0 represents a hospital, 1 represents a school, and 2 represents a company.

[0028] Step S202: Input the comprehensive attribute information into the trained sensitive user classification model to obtain the sensitive user classification result corresponding to the user to be identified.

[0029] The sensitive user classification model can adopt a triplet Siamese network model, and the sensitive user classification result can refer to the sensitive user classification category predicted based on the comprehensive attribute information of the user to be identified.

[0030] Specifically, the sensitive user classification model can adopt a single-branch structure. In this case, the architecture of the sensitive user classification model can include an encoder and a distance comparator. The encoder can be implemented using convolutional layers, multi-head attention layers, etc. The encoder takes the vector corresponding to the comprehensive attribute information as input and outputs the target feature vector corresponding to the comprehensive attribute information. The distance comparator is used to measure the distance between the target feature vector and the pre-acquired first reference feature vector and second reference feature vector respectively to obtain the first distance and the second distance. Then, the first distance and the second distance are compared, and the category to which the reference feature vector corresponding to the smaller value belongs is taken as the sensitive user classification result corresponding to the user to be identified. The first reference feature vector can be obtained by inputting the comprehensive attribute information of random positive samples, that is, random sensitive users, into the encoder. The second reference feature vector can be obtained by inputting the comprehensive attribute information of random negative samples, that is, random non-sensitive users, into the encoder.

[0031] In one embodiment, the first reference feature vector can also be determined based on the average of the feature vectors obtained by inputting the comprehensive attribute information of a batch of sensitive users into the encoder, and the second reference feature vector can also be determined based on the average of the feature vectors obtained by inputting the comprehensive attribute information of a batch of non-sensitive users into the encoder.

[0032] In one specific implementation, the power grid sensitive user identification method further includes:

[0033] Obtain the basic attribute information, complaint information, and power supply sequence corresponding to M power grid users with complaint behavior, where M is a positive integer, and the basic attribute information includes user type;

[0034] For any power grid user with a complaint, the reference time interval for that power grid user is determined based on the reference time point corresponding to the complaint information and the preset time deviation value.

[0035] By blocking the power supply corresponding to the time point covered by the reference time interval in the power supply sequence of the power supply user of the power grid, the blocking sequence corresponding to the power grid user is obtained.

[0036] Based on the user type corresponding to the power grid user, determine the power supply time series prediction model corresponding to the power grid user;

[0037] Input the shielding sequence corresponding to the power grid user into the power supply time series prediction model corresponding to the power grid user to obtain the prediction sequence corresponding to the power grid user;

[0038] Based on the power supply sequence and forecast sequence corresponding to the power grid user, determine the initial sensitivity evaluation value corresponding to the power grid user;

[0039] Based on the initial sensitivity evaluation value, user type, and complaint information corresponding to the power grid user, determine the target sensitivity evaluation value corresponding to the power grid user;

[0040] Based on the target sensitivity evaluation values ​​corresponding to the M power grid users with complaint behavior, determine the evaluation value threshold;

[0041] Based on the evaluation threshold and the target sensitivity evaluation values ​​corresponding to the M power grid users with complaint behavior, the M power grid users are divided into A positive samples and B negative samples, where A+B=M;

[0042] Based on the comprehensive attribute information corresponding to each positive sample and the comprehensive attribute information corresponding to each negative sample, a sensitive user classification model is trained to obtain a well-trained sensitive user classification model.

[0043] Where M can refer to the total number of users with complaint behavior, the reference time point T can refer to the time point corresponding to a single complaint record, the time deviation value can refer to the preset time range parameter t, the time range parameter can be set to 2 hours in this embodiment, and the implementer can adjust the time deviation value according to the actual situation. The reference time interval can be represented as [Tt, T+t].

[0044] Blocking the power supply corresponding to the time point covered by the reference time interval in the power supply sequence of the power supply user of the power grid can mean setting the power supply corresponding to the time point in the power supply sequence that belongs to the reference time interval to a preset value, which can be 0.

[0045] The power supply time series prediction model can adopt long short-term memory network model, temporal convolution model, etc. The structure and training method of the power supply time series prediction model will not be described in detail here. However, it should be noted that since the power consumption habits of different user types usually have great differences, this embodiment pre-sets a corresponding power supply time series prediction model for each user type. It is only necessary to select the power supply sequence corresponding to the user of the same user type as the training sample when preparing the training sample to train the power supply time series prediction model.

[0046] The initial sensitivity evaluation value can represent the normalized quantitative value of the deviation between the actual power supply fluctuation and the model prediction during the complaint period. The larger the initial sensitivity evaluation value, the smaller the sensitivity of the corresponding power grid user, and vice versa.

[0047] The target sensitivity evaluation value can refer to the sensitivity of power grid users after adjusting for user type and complaint information.

[0048] Evaluation thresholds can be used to distinguish between sensitive and non-sensitive users among power grid users who have complained.

[0049] Positive samples can refer to power grid users with complaint behavior whose target sensitivity evaluation value is less than the evaluation value threshold. Positive samples represent sensitive users. Negative samples can refer to power grid users with complaint behavior whose target sensitivity evaluation value is greater than or equal to the evaluation value threshold. Negative samples represent non-sensitive users.

[0050] In one specific implementation, the power grid sensitive user identification method further includes:

[0051] Obtain N power grid users who have not filed any complaints as negative samples, where N is a positive integer.

[0052] In this embodiment, non-complaining users are added as negative samples, and the model's generalization ability is poor because the encoding is trained only with complaining users.

[0053] In one specific implementation, the initial sensitivity assessment value for the power grid user is determined based on the power supply sequence and the predicted sequence, including:

[0054] Calculate the Euclidean distance between the power supply sequence and the predicted sequence corresponding to the power grid user to obtain the first distance;

[0055] The first distance is compared with the preset reference distance, and the ratio is used as the initial sensitivity evaluation value for the power grid user. The preset reference distance is the maximum value of the first distance for each of the M power grid users with complaint behavior.

[0056] The first distance can be used to measure the difference between the power supply sequence and the predicted sequence, while the preset reference distance can be used for normalization.

[0057] In one implementation, the implementer may also use other distance metrics to measure the difference between the power supply sequence and the predicted sequence, such as cosine similarity.

[0058] In one specific implementation, the complaint information includes the number of complaints;

[0059] Based on the initial sensitivity assessment value, user type, and complaint information corresponding to the power grid user, determine the target sensitivity assessment value corresponding to the power grid user, including:

[0060] Based on the number of complaints corresponding to the power grid users, the first adjustment coefficient is obtained by mapping through a preset mapping function;

[0061] The second regulation coefficient is determined by looking up a table based on the user type corresponding to the power grid user.

[0062] The initial sensitivity evaluation value corresponding to the power grid user is multiplied by the first adjustment coefficient and the second adjustment coefficient, and the result of the multiplication is used as the target sensitivity evaluation value corresponding to the power grid user.

[0063] The preset mapping function maps the number of complaints I to a first adjustment coefficient a, and the preset mapping function can be expressed as a = -(I-1). 2 / (max-1) 2 +1, where max is the maximum number of complaints from all users. It can be seen that the larger I is, the smaller a is.

[0064] The lookup table can include the correspondence between user type and second adjustment coefficient. In this embodiment, the second adjustment coefficient b=0.2 for hospital type, b=0.3 for school type, b=0.5 for enterprise type, and b=1 for family type.

[0065] Specifically, the smaller the sensitivity evaluation value, the greater the sensitivity of the corresponding power grid user. Therefore, when there are many user complaints or the user belongs to a special user type, the user's initial sensitivity evaluation value is corrected to be smaller by the first adjustment coefficient and the second adjustment coefficient to avoid misjudgment.

[0066] In one specific implementation, an evaluation threshold is determined based on the target sensitivity evaluation values ​​corresponding to the M power grid users with complaint behavior, including:

[0067] Based on the target sensitivity evaluation values ​​corresponding to the M power grid users with complaint behavior, a preset clustering algorithm is used to perform clustering processing to obtain K clustering results, where K=2;

[0068] The evaluation threshold is determined based on the evaluation value ranges corresponding to the two clustering results.

[0069] The preset clustering algorithm can be the K-means clustering algorithm. The number of cluster sets of the K-means clustering algorithm is set to 2. That is, based on the target sensitivity evaluation value corresponding to the M power grid users with complaint behavior, the M power grid users are divided into high sensitivity clusters and low sensitivity clusters.

[0070] The evaluation value range can refer to the minimum range of target sensitivity evaluation values ​​that include all power grid users in the corresponding clustering results.

[0071] Specifically, the evaluation value threshold is determined based on the evaluation value ranges corresponding to the two clustering results. This can be achieved by averaging the maximum value of the evaluation value range of the high-sensitivity cluster and the minimum value of the evaluation value range of the low-sensitivity cluster, and then using the average value as the evaluation value threshold.

[0072] Step S203: When the sensitive user classification result corresponding to the user to be identified meets the preset conditions, the user to be identified is determined to be a power grid sensitive user.

[0073] Among them, the preset conditions are used to determine whether the user to be identified is a grid-sensitive user. When it is determined that the user to be identified is a grid-sensitive user, the grid management side can optimize the power supply strategy for the grid-sensitive user in a targeted manner to reduce the complaint rate and improve the user's electricity experience.

[0074] In one specific implementation, the sensitive user classification results include sensitive user categories and non-sensitive user categories;

[0075] Accordingly, the preset condition is: the sensitive user classification result corresponding to the user to be identified is a sensitive user category.

[0076] Among them, the sensitive user category can refer to the user category that is highly sensitive to power supply quality, and that even slight abnormal fluctuations in power supply quality will cause complaints or significant losses. Power supply quality can refer to voltage stability and power supply continuity.

[0077] Non-sensitive user category can refer to user category that has a high tolerance for power quality fluctuations and is not likely to complain due to minor abnormalities.

[0078] In this embodiment, by fusing multi-dimensional data such as basic attributes, electricity consumption behavior, complaint information, and service information, the one-sidedness of identifying sensitive users of the power grid based on single data is avoided, thus improving the reliability of identifying sensitive users of the power grid. The sensitive user classification model realizes automated user identification without the need for manual intervention in the identification process, thereby improving the efficiency of identifying sensitive users of the power grid. Furthermore, it can identify sensitive users through multi-dimensional data even without complaint information, avoiding the lag in identifying some sensitive users that can only be completed after a complaint is received. This effectively reduces the complaint rate and user losses, and improves the precision of power grid user management.

[0079] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0080] In one embodiment, a method and apparatus for identifying power grid sensitive users based on multi-dimensional data perception is provided, which corresponds one-to-one with the power grid sensitive user identification method based on multi-dimensional data perception in the above embodiments. For example... Figure 3 As shown, the functional modules of the device for identifying power grid sensitive users based on multi-dimensional data perception are described in detail below:

[0081] The multi-dimensional information acquisition module 301 is used to acquire comprehensive attribute information of the user to be identified, wherein the comprehensive attribute information includes at least basic attribute information, electricity consumption behavior information, historical complaint information and historical service information;

[0082] The model inference module 302 is used to input comprehensive attribute information into the trained sensitive user classification model to obtain the sensitive user classification result corresponding to the user to be identified;

[0083] The sensitive user identification module 303 is used to identify the user to be identified as a power grid sensitive user when the sensitive user classification result corresponding to the user to be identified meets the preset conditions.

[0084] Specific limitations regarding the grid-sensitive user identification method and device based on multi-dimensional data sensing can be found in the limitations of the grid-sensitive user identification method based on multi-dimensional data sensing mentioned above, and will not be repeated here. Each module in the aforementioned grid-sensitive user identification method and device based on multi-dimensional data sensing can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0085] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 4 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for identifying power grid-sensitive users based on multi-dimensional data sensing.

[0086] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the power grid sensitive user identification method based on multi-dimensional data perception described in the above embodiment. To avoid repetition, this will not be repeated here. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in this embodiment of the power grid sensitive user identification method apparatus based on multi-dimensional data perception. To avoid repetition, this will not be repeated here.

[0087] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When executed by a processor, the computer program implements the grid-sensitive user identification method based on multi-dimensional data perception described in the above embodiment. To avoid repetition, this will not be described again here. Alternatively, when executed by a processor, the computer program implements the functions of each module / unit in this embodiment of the grid-sensitive user identification method apparatus based on multi-dimensional data perception. To avoid repetition, this will not be described again here.

[0088] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0089] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0090] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for identification of sensitive consumers of power grid based on multi-dimensional data perception, characterized in that, The power grid sensitive user identification method comprises: obtaining comprehensive attribute information of a user to be identified, wherein the comprehensive attribute information at least comprises basic attribute information, power consumption behavior information, historical complaint information and historical service information; inputting the comprehensive attribute information into a trained sensitive user classification model to obtain a sensitive user classification result corresponding to the user to be identified; when the sensitive user classification result corresponding to the user to be identified meets a preset condition, determining that the user to be identified is a power grid sensitive user.

2. The method of grid sensitive customer identification as claimed in claim 1, wherein, The sensitive user classification result comprises a sensitive user category and a non-sensitive user category; Correspondingly, the preset condition is that the sensitive user classification result corresponding to the user to be identified is the sensitive user category.

3. The method of grid sensitive customer identification as claimed in claim 1 wherein, The power grid sensitive user identification method further comprises: obtaining basic attribute information, complaint information and power supply quantity sequence corresponding to M power grid users with complaint behaviors, wherein M is a positive integer, and the basic attribute information comprises user types; for any power grid user with complaint behaviors, determining a reference time interval corresponding to the power grid user according to a reference time point corresponding to the complaint information of the power grid user and a preset time deviation value; obtaining a shielding sequence corresponding to the power grid user by shielding power supply quantity corresponding to time points covered by the reference time interval in the power supply quantity sequence of the power grid user; determining a power supply quantity time sequence prediction model corresponding to the power grid user according to the user type corresponding to the power grid user; inputting the shielding sequence corresponding to the power grid user into the power supply quantity time sequence prediction model corresponding to the power grid user to obtain a prediction sequence corresponding to the power grid user; determining an initial sensitive degree evaluation value corresponding to the power grid user according to the power supply quantity sequence and the prediction sequence corresponding to the power grid user; determining a target sensitive degree evaluation value corresponding to the power grid user according to the initial sensitive degree evaluation value, the user type and the complaint information corresponding to the power grid user; determining an evaluation value threshold according to the target sensitive degree evaluation values corresponding to the M power grid users with complaint behaviors; dividing the M power grid users into A positive samples and B negative samples according to the evaluation value threshold and the target sensitive degree evaluation values corresponding to the M power grid users with complaint behaviors, wherein A+B=M; training the sensitive user classification model according to the comprehensive attribute information corresponding to each positive sample and the comprehensive attribute information corresponding to each negative sample to obtain the trained sensitive user classification model.

4. The method of grid sensitive customer identification of claim 3, wherein, The power grid sensitive user identification method further comprises: obtaining N power grid users without complaint behaviors as negative samples, wherein N is a positive integer.

5. The method of grid sensitive customer identification as claimed in claim 3, wherein, The determination of the initial sensitive degree evaluation value corresponding to the power grid user according to the power supply quantity sequence and the prediction sequence corresponding to the power grid user comprises: calculating the Euclidean distance between the power supply quantity sequence and the prediction sequence corresponding to the power grid user to obtain a first distance; comparing the first distance with a preset reference distance, and taking the ratio as the initial sensitive degree evaluation value corresponding to the power grid user, wherein the preset reference distance is the maximum value of the first distances corresponding to the M power grid users with complaint behaviors.

6. The method of grid sensitive customer identification as claimed in claim 3, wherein, The complaint information comprises the number of complaints. The target sensitive degree evaluation value corresponding to the power grid user is determined according to the initial sensitive degree evaluation value, the user type and the complaint information of the power grid user, and the target sensitive degree evaluation value corresponding to the power grid user is determined. A first adjustment coefficient is obtained by mapping a preset mapping function according to the number of complaints corresponding to the power grid user; A second adjustment coefficient is determined by looking up a table according to the user type corresponding to the power grid user; The initial sensitive degree evaluation value corresponding to the power grid user is multiplied by the first adjustment coefficient and the second adjustment coefficient, and the multiplication result is taken as the target sensitive degree evaluation value corresponding to the power grid user.

7. The method of grid sensitive customer identification of claim 6, wherein, The evaluation value threshold is determined according to the target sensitive degree evaluation value corresponding to the M power grid users with complaint behaviors. The target sensitive degree evaluation value corresponding to the M power grid users with complaint behaviors is clustered by using a preset clustering algorithm to obtain K clustering results, wherein K = 2. The evaluation value threshold is determined according to the evaluation value range corresponding to the two clustering results.

8. A device for identification of sensitive consumers of power grid based on multi-dimensional data perception, characterized by The power grid sensitive user identification device comprises: A multi-dimensional information acquisition module is configured to acquire comprehensive attribute information of a user to be identified, wherein the comprehensive attribute information at least includes basic attribute information, power consumption behavior information, historical complaint information and historical service information; A model reasoning module is configured to input the comprehensive attribute information into a trained sensitive user classification model to obtain a sensitive user classification result corresponding to the user to be identified; A sensitive user identification module is configured to determine that the user to be identified is a power grid sensitive user when the sensitive user classification result corresponding to the user to be identified meets a preset condition.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the power grid sensitive user identification method based on multi-dimensional data perception according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the power grid sensitive user identification method based on multi-dimensional data perception according to any one of claims 1 to 7.

Images