A power enterprise marketing account intelligent checking supervision method
By analyzing users' historical electricity consumption and electricity bill characteristic sequences, identifying suspected abnormal data points and calculating accounting risk coefficients, intelligent verification of power companies' marketing accounts has been achieved, improving the accuracy and efficiency of verification and reducing the risk of user complaints.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 国网新疆电力有限公司营销服务中心
- Filing Date
- 2026-03-23
- Publication Date
- 2026-06-09
Smart Images
Figure CN122175722A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of data monitoring technology, specifically to a method for intelligent verification and supervision of marketing accounts in power companies. Background Technology
[0002] Power companies conduct marketing account reconciliation to ensure the accuracy of electricity billing, improve customer service, and reduce customer complaints. Marketing account reconciliation involves cross-checking data from multiple stages, including metering data, billing rules, electricity pricing policies, and billing results. This can identify inconsistencies in electricity consumption and billing, as well as errors in the application of rates, thus preventing accounting errors from flowing into the settlement and billing process and reducing operational risks.
[0003] Currently, power companies typically use manual verification methods, randomly checking a subset of users within a target scope. However, this randomness leads to low accuracy in supervision. Furthermore, verifying only electricity consumption and charges results in poor early warning capabilities for potential accounting risks. Anomalies in accounts can lead to continuous error correction, impacting reconciliation efficiency. Additionally, if accounting anomalies occur at the user's side, it can easily lead to complaints, re-inspections, and decreased trust. Therefore, how to efficiently and specifically verify the accounts of users with potential accounting risks has become an urgent problem to be solved. Summary of the Invention
[0004] To address the aforementioned technical problems, the present invention aims to provide an intelligent verification and supervision method for marketing accounts in power enterprises. The specific technical solution adopted is as follows:
[0005] Obtain two distinct electricity consumption behavior feature sequences from the user's historical daily electricity consumption and daily electricity cost;
[0006] Based on the data discreteness and local data fluctuation characteristics of the electricity consumption behavior feature sequence, the degree of suspected abnormal fluctuation of data points is obtained; based on the degree of suspected abnormal fluctuation, suspected abnormal data points in the electricity consumption behavior feature sequence are obtained; based on the data distribution characteristics of the suspected abnormal data points, similar data points are obtained; based on the quantity characteristics of similar data points, the data difference characteristics of similar data points, and the time interval characteristics, the degree of significant fluctuation of suspected abnormal data points is obtained.
[0007] The degree of significant anomaly of the suspected abnormal data points is obtained based on the differences in the degree of significant fluctuations at the same position in the suspected abnormal data points and other electricity consumption behavior characteristic sequences, as well as the differences in local data changes; the user's billing risk coefficient is obtained based on the number of suspected abnormal data points, the degree of significant anomaly, and the time interval from the current date in the two electricity consumption behavior characteristic sequences.
[0008] The electricity bill monitoring range is adaptively adjusted based on the aforementioned accounting risk coefficient to obtain the user's adaptive monitoring range; accounting reconciliation and monitoring are then performed based on the aforementioned adaptive monitoring range.
[0009] Furthermore, the step of obtaining the degree of suspected abnormal fluctuations of data points based on the data discrete characteristics and local data fluctuation characteristics of the electricity consumption behavior characteristic sequence includes:
[0010] In the formula This indicates the degree of suspected abnormal fluctuations at the i-th data point. Indicates normalization, Let represent the value of the i-th data point, and H represent the average value of the electricity consumption behavior characteristic sequence. Indicates the first The value of each data point. express The value of each data point.
[0011] Further, the step of obtaining suspected abnormal data points in the electricity consumption behavior feature sequence based on the suspected abnormal fluctuation degree includes:
[0012] Data points in the electricity consumption behavior feature sequence whose suspected abnormal fluctuations exceed a preset threshold are considered as suspected abnormal data points.
[0013] Furthermore, the step of obtaining similar data points based on the data distribution characteristics of the suspected abnormal data points includes:
[0014] Based on the values of the suspected abnormal data points, the suspected abnormal data points are clustered to obtain different clusters; suspected abnormal data points within the same cluster are considered as similar data points.
[0015] Furthermore, the step of obtaining the degree of significant fluctuation of the suspected abnormal data points based on the quantity characteristics of similar data points, the data difference characteristics of similar data points, and the time interval characteristics includes:
[0016] In the formula This indicates the degree of significant fluctuation of the p-th suspected outlier data point. Let N represent an exponential function with the natural constant as the base, and let N represent the number of similar data points to the suspected outlier data point. Let F represent the value of the nth similar data point, and let F represent the average value of the similar data points of the suspected abnormal data point. This indicates the number of intervals between adjacent similar data points. G represents the length of the interval between the m-th adjacent data points, and G represents the average interval length between adjacent similar data points.
[0017] Further, the step of obtaining the significant anomaly degree of the suspected abnormal data point based on the difference characteristics of significant fluctuations at the same position in the suspected abnormal data point and other electricity consumption behavior characteristic sequences, and the difference characteristics of local data changes, includes:
[0018] In the formula This indicates the degree of significant anomaly of the p-th suspected outlier data point. This indicates the degree of significant fluctuation in suspected outlier data points. This indicates the degree of significant fluctuation of suspected abnormal data points at the same position in other electricity consumption behavior characteristic sequences. This represents the slope of the change between the p-th suspected outlier data point and the previous data point. This represents the slope of the change from the previous data point at the same position in other electricity consumption behavior characteristic sequences.
[0019] Furthermore, the step of obtaining the user's billing risk coefficient based on the quantity characteristics, degree of significant anomaly, and time interval characteristics of suspected abnormal data points in the two electricity consumption behavior characteristic sequences includes:
[0020] In the formula, E represents the user's billing risk coefficient, Q represents the number of electricity consumption behavior feature sequences, and U represents the number of suspected abnormal data points in the electricity consumption behavior feature sequences. This indicates the degree of significant anomaly of the p-th suspected outlier data point. This represents the time interval between the p-th suspected abnormal data point and the current date.
[0021] Furthermore, the step of adaptively adjusting the electricity bill monitoring interval based on the accounting risk coefficient to obtain the user's adaptive monitoring interval includes:
[0022] , In the formula This indicates the adaptive minimum electricity cost. Indicates the adaptive maximum electricity price, 'a' represents the preset minimum electricity price, 'b' represents the preset maximum electricity price, and 'E' represents the accounting risk coefficient. and Construct adaptive monitoring intervals for users.
[0023] The present invention has the following beneficial effects:
[0024] In this invention, obtaining the degree of suspected abnormal fluctuations can identify data points in the electricity consumption behavior feature sequence that exhibit outliers and sudden increases or decreases. Obtaining suspected abnormal data points can identify relatively abnormal electricity consumption behaviors in a user's historical electricity consumption data, thereby determining whether the user has potential electricity billing anomalies. Obtaining similar data points can identify similar electricity consumption behaviors in the user's history. Obtaining the degree of significant fluctuations allows for further analysis of suspected abnormal data points, thereby determining whether they represent regular electricity consumption fluctuations. Obtaining the degree of significant anomalies, combined with the user's daily electricity consumption characteristics and daily electricity bill characteristics, can more accurately determine whether the user has potential electricity billing anomalies. Based on the quantity characteristics, degree of significant anomalies, and time interval characteristics of suspected abnormal data points in the two electricity consumption behavior feature sequences, a user's billing risk coefficient is obtained, which can characterize the degree of potential billing anomalies and achieve billing risk early warning. Finally, the electricity billing monitoring interval is adaptively adjusted based on the billing risk coefficient to obtain an adaptive monitoring interval for the user, enabling staff to specifically extract and calculate billing for users who may have potential billing anomalies, improving the accuracy of electricity bill settlement. Attached Figure Description
[0025] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art 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.
[0026] Figure 1 The flowchart illustrates a method for intelligent verification and supervision of marketing accounts in power companies, as provided in one embodiment of the present invention. Detailed Implementation
[0027] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a smart verification and supervision method for marketing accounts of power enterprises proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.
[0028] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.
[0029] The following description, in conjunction with the accompanying drawings, details a specific scheme for an intelligent verification and supervision method for marketing accounts of power enterprises provided by this invention.
[0030] Please see Figure 1 The diagram illustrates a flowchart of an intelligent verification and supervision method for marketing accounts of power companies, provided by an embodiment of the present invention. The method includes the following steps:
[0031] Step S1: Obtain two different electricity consumption behavior feature sequences: the user's historical daily electricity consumption and daily electricity bill.
[0032] By collecting daily electricity consumption data and daily electricity charges from the power system, a file is created for each electricity user. A daily electricity consumption sequence and a daily electricity charge sequence are constructed based on the user's daily electricity consumption data and daily electricity charge data, respectively. The daily electricity consumption data and daily electricity charge data in the two sequences are normalized to obtain two distinct historical electricity consumption behavior characteristic sequences for the user. Normalizing the daily electricity consumption and daily electricity charge data can avoid the influence of dimensions in subsequent analysis. Implementers can determine the collection cycle of user behavior characteristic sequences according to the implementation scenario.
[0033] Step S2: Obtain the degree of suspected abnormal fluctuation of data points based on the data discrete characteristics and local data fluctuation characteristics of the electricity consumption behavior feature sequence; obtain suspected abnormal data points in the electricity consumption behavior feature sequence based on the degree of suspected abnormal fluctuation; obtain similar data points based on the data distribution characteristics of suspected abnormal data points; obtain the degree of significant fluctuation of suspected abnormal data points based on the quantity characteristics of similar data points, the data difference characteristics of similar data points, and the time interval characteristics.
[0034] Potential anomalies in user electricity billing typically manifest as unusual fluctuations in electricity consumption or billing data, such as sudden spikes or drops. To assess historical electricity consumption behavior data and construct a user risk profile, the first step is to identify potential anomaly fluctuation points within the electricity consumption behavior feature sequence. When a value in the electricity consumption behavior feature sequence deviates significantly from the normal value, it indicates an outlier in the user's electricity consumption or billing for that day. Therefore, the historical average can be used as a benchmark. The greater the difference between a data point and the historical average, the more outlier the data point is compared to other data points, and the more likely it is to be an outlier. Conversely, when the difference between a data point and its adjacent data points is too large, it indicates a significant spike or drop in the data, and the user's electricity consumption behavior on that day does not conform to the data changes of adjacent days, demonstrating obvious anomaly fluctuations. Therefore, the degree of suspected anomaly fluctuation for a data point can be obtained based on the data dispersion characteristics and local data fluctuation characteristics of the electricity consumption behavior feature sequence. A data point refers to the electricity consumption feature data of any day within the electricity consumption behavior feature sequence. Preferably, in this embodiment of the invention, the step of obtaining the degree of suspected anomaly fluctuation for a data point includes:
[0035]
[0036] In the formula, This indicates the degree of suspected abnormal fluctuations at the i-th data point. This indicates normalization, which can normalize the maximum and minimum values based on the calculation results of all the same formulas for that user. Let represent the value of the i-th data point, and H represent the average value of the electricity consumption behavior characteristic sequence. Indicates the first The value of each data point. express The value of each data point. When The larger the value, the more outlier the data point is compared to other data points, and the stronger the suspected anomalous fluctuation of the data point. When The larger the value, the more pronounced the sudden rises and falls in the data point, and the stronger the suspected anomalous fluctuation. Therefore, the stronger the suspected anomalous fluctuation of a data point, the more obvious its outlier and sudden rise / fall characteristics. It should be noted that the suspected anomalous fluctuation is calculated separately for each of the user's two electricity consumption behavior characteristic sequences.
[0037] Furthermore, after obtaining the degree of suspected abnormal fluctuations of different data points in the user's electricity consumption behavior feature sequence, suspected abnormal data points in the electricity consumption behavior feature sequence can be obtained based on the degree of suspected abnormal fluctuations. Preferably, in this embodiment of the invention, the step of obtaining suspected abnormal data points includes: data points in the electricity consumption behavior feature sequence whose degree of suspected abnormal fluctuations exceeds a preset threshold as suspected abnormal data points. Since the value of the degree of suspected abnormal fluctuations after normalization is between 0 and 2, the degree of suspected abnormal fluctuations of data points with obvious abnormal fluctuations is close to 2, while the degree of suspected abnormal fluctuations of data points without fluctuations is close to 0. The two are located at opposite ends of the value range. Therefore, in this embodiment of the invention, the preset threshold is 1, which can be determined by the implementer according to the implementation scenario. Suspected anomaly data points represent data points in the electricity consumption behavior sequence that are relatively outlier and exhibit sudden increases or decreases. Since fluctuations in electricity consumption or bills in actual user situations are influenced by various factors—for example, electricity consumption in office or residential areas is affected by holidays, and factory electricity consumption is affected by production schedules—unusual electricity consumption and costs may occur in real-world scenarios. Therefore, volatile data cannot be directly identified as data representing potential electricity billing risks; further assessment of the degree of significant fluctuation is required. The more data points with similar fluctuation characteristics to suspected anomaly data points exist at other historical time points, and the higher the degree of similarity, the more likely similar electricity consumption behaviors have occurred historically. These data points may represent normal electricity consumption behavior changes with significant fluctuations on that day. Furthermore, the more similar the time intervals between similar data points, the more regular the electricity consumption behavior is, and the less it conforms to the random and sudden characteristics of electricity billing anomalies. Therefore, similar data points are obtained based on the data distribution characteristics of suspected abnormal data points. Preferably, in this embodiment of the invention, the step of obtaining similar data points includes: clustering suspected abnormal data points according to their values to obtain different clusters; the values of suspected abnormal data points within a cluster are similar, indicating similar electricity consumption behavior on that day; it should be noted that the clustering object is suspected abnormal data points in the same electricity consumption behavior feature sequence; therefore, suspected abnormal data points within the same cluster are considered similar data points. Furthermore, the significant fluctuation degree of suspected abnormal data points can be obtained based on the quantity characteristics of similar data points, the data difference characteristics of similar data points, and the time interval characteristics; preferably, in this embodiment of the invention, the step of obtaining the significant fluctuation degree includes:
[0038]
[0039] In the formula, This indicates the degree of significant fluctuation of the p-th suspected outlier data point. The exponential function is defined with the natural constant as its base. N represents the number of similar data points to the suspected outlier, which is the number of data points in the cluster to which the suspected outlier belongs. The larger this number is, the more dates there are with similar electricity consumption behavior, and the more likely it is to be a real fluctuation in electricity consumption. Let F represent the value of the nth similar data point, and let F represent the average value of the similar data points of the suspected outlier. The smaller the value, the more similar the values of similar data points are, and the more likely it is to be a real fluctuation in electricity consumption. This indicates the number of intervals between adjacent similar data points. Let G represent the length of the interval between the m-th adjacent data points, and let G represent the average interval length between adjacent similar data points. The time interval is in days. The smaller the value, the more similar the time intervals between similar data points, indicating a higher likelihood of regular and similar electricity consumption behavior, and thus a greater likelihood of genuine electricity consumption fluctuations. Therefore, the greater the significant fluctuation of a suspected abnormal data point, the more likely its fluctuation represents unusual electricity consumption behavior fluctuations by the user, and the more likely the user has potential electricity billing risks. It should be noted that if the denominator is 0, a preset minimum positive number is used instead of the denominator for calculation. In this embodiment of the invention, the preset minimum positive number is 0.01.
[0040] Step S3: Obtain the significant abnormality of the suspected abnormal data points based on the differences in the degree of significant fluctuations at the same location in the suspected abnormal data points and other electricity consumption behavior characteristic sequences, as well as the differences in local data changes; obtain the user's billing risk coefficient based on the number of suspected abnormal data points, the degree of significant abnormality, and the time interval from the current date in the two electricity consumption behavior characteristic sequences.
[0041] Because analyzing abnormal fluctuations in daily electricity consumption and electricity cost alone has limitations, it's possible for electricity consumption to fluctuate abnormally while electricity costs remain within the normal range, or vice versa. Therefore, it's possible to analyze the difference in significant fluctuations between suspected abnormal data points in one dimension and data points in the same time period in another dimension. The larger the difference, the more consistent it is with the situation where electricity consumption fluctuates abnormally while electricity costs remain within the normal range, or vice versa, thus indicating a genuine anomaly in the suspected data point. However, simultaneous anomalies in both electricity consumption and electricity costs do not necessarily prove that the suspected data point is normal. Further analysis of the changing trend characteristics of the two dimensions is needed. If the changing trends are the same, the more the suspected data point conforms to the pattern of electricity costs changing with electricity consumption, and the lower the degree of genuine anomaly. Therefore, the significant anomaly degree of a suspected abnormal data point can be obtained based on the difference in significant fluctuations at the same position in other electricity behavior feature sequences and the difference in local data changes. Preferably, in this embodiment of the invention, the step of obtaining the significant anomaly degree includes:
[0042]
[0043] In the formula, This indicates the degree of significant anomaly of the p-th suspected outlier data point. This indicates the degree of significant fluctuation in suspected outlier data points. This indicates the degree of significant fluctuation of suspected outlier data points at the same location in other electricity consumption behavior characteristic sequences. If no suspected outlier data points are found at the same location in other electricity consumption behavior characteristic sequences, then... It is 0. This represents the slope of the change between the p-th suspected outlier data point and the previous data point. This represents the slope of the change from the previous data point at the same position in other electricity consumption behavior characteristic sequences. The slope is the slope of the line connecting two points in a coordinate system representing the electricity consumption behavior characteristic sequence. When... The larger the value, the lower the similarity of the significant fluctuations of the two electricity consumption behavior characteristic sequences at the same time, and the more likely there is a risk of abnormal electricity billing. The greater the degree of significant abnormality of the suspected abnormal data point; when... The smaller the value, the more similar the slope of change, indicating a more similar trend in electricity consumption and its changes, and a greater likelihood that the electricity bill reflects a true change in electricity consumption. Conversely, a larger value suggests a greater likelihood that the trends in electricity bill and electricity consumption are inconsistent. Therefore, a higher degree of significant anomaly indicates a greater risk of abnormal electricity billing for the user. During the calculation process, the results of all identical calculation formulas for the user are normalized to their maximum and minimum values.
[0044] Furthermore, the more suspected abnormal data points there are in the user's historical electricity consumption data, and the greater the degree of significant abnormality of each suspected abnormal data point, the higher the user's electricity billing risk level, and the more necessary it is for staff to conduct targeted verification of the user's electricity consumption data and electricity bills. Therefore, the user's billing risk coefficient can be obtained based on the quantity characteristics, degree of significant abnormality, and time interval characteristics of suspected abnormal data points in two electricity consumption behavior feature sequences; preferably, in this embodiment of the invention, the step of obtaining the billing risk coefficient includes:
[0045]
[0046] In the formula, E represents the user's billing risk coefficient, Q represents the number of electricity consumption behavior feature sequences, and U represents the number of suspected abnormal data points in the electricity consumption behavior feature sequences. This indicates the degree of significant anomaly of the p-th suspected outlier data point. This represents the time interval between the p-th suspected abnormal data point and the current date. The closer the time interval, the greater its impact on the current real-time billing risk assessment; conversely, the farther away the time interval, the smaller its impact on the current assessment. Therefore, the more suspected abnormal data points a user has in two electricity consumption behavior characteristic sequences, the greater the degree of significant abnormality, and the closer the time interval is to the current date, the higher the user's billing risk coefficient. This indicates that the user is more likely to have potential electricity billing abnormalities and requires more targeted verification by staff. This normalization uses the same calculation formula for all users within the same power area to normalize the maximum and minimum values.
[0047] Step S4: Adaptively adjust the electricity bill monitoring range according to the accounting risk coefficient to obtain the user's adaptive monitoring range; conduct accounting reconciliation monitoring based on the adaptive monitoring range.
[0048] Users' daily electricity bills are usually within a normal range. When they exceed this range, the user needs to be closely monitored. Therefore, if a user's accounting risk coefficient is high, the user's electricity bill monitoring range can be adjusted to make the user more easily noticed by staff, thus allowing for targeted accounting. Therefore, the electricity bill monitoring range can be adaptively adjusted based on the accounting risk coefficient to obtain the user's adaptive monitoring range. Preferably, in this embodiment of the invention, the step of obtaining the adaptive monitoring range includes: , In the formula This indicates the adaptive minimum electricity cost. Indicates the adaptive maximum electricity price, 'a' represents the preset minimum electricity price, 'b' represents the preset maximum electricity price, and 'E' represents the accounting risk coefficient. and Establish an adaptive monitoring range for each user. The higher a user's accounting risk coefficient, the smaller the adaptive monitoring range becomes, making it easier for the user's electricity bills to exceed the range and triggering an early warning, thus identifying potential accounting risks as early as possible. It should be noted that the preset minimum and maximum electricity charges can be determined by the implementer based on the implementation scenario.
[0049] Furthermore, accounting reconciliation and monitoring can be conducted based on adaptive monitoring intervals. When a user's daily electricity bill exceeds the adaptive monitoring interval, staff can further calculate the user's usage and electricity charges, thereby identifying potential accounting risks as early as possible and preventing data anomalies that could affect the rights and interests of both the power company and the user. Thus, adaptive monitoring intervals can help power company staff efficiently and specifically calculate user electricity billing issues, improving the accuracy of electricity bill settlement.
[0050] In summary, this invention provides an intelligent verification and supervision method for power company marketing accounts. It identifies suspected abnormal data points based on the discrete characteristics and local data fluctuation characteristics of electricity consumption behavior feature sequences. It then determines the degree of significant fluctuation based on the quantity, data difference, and time interval characteristics of similar data points to the suspected abnormal data points. Finally, it determines the degree of significant anomaly based on the differences in the degree of significant fluctuation and local data changes between the suspected abnormal data points and other electricity consumption behavior feature sequences at the same location. Finally, it obtains the user's account risk coefficient based on the quantity, degree of significant anomaly, and time interval from the current date of the suspected abnormal data points. This invention adaptively adjusts the electricity bill monitoring interval based on the account risk coefficient and performs account verification monitoring based on the adaptive monitoring interval, making electricity bill verification more efficient and targeted, and improving the accuracy of electricity bill settlement.
[0051] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0052] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.
Claims
1. A method for intelligent verification and supervision of marketing accounts in power enterprises, characterized in that, The method includes the following steps: Obtain two distinct electricity consumption behavior feature sequences from the user's historical daily electricity consumption and daily electricity cost; Based on the data discreteness and local data fluctuation characteristics of the electricity consumption behavior feature sequence, the degree of suspected abnormal fluctuation of data points is obtained; based on the degree of suspected abnormal fluctuation, suspected abnormal data points in the electricity consumption behavior feature sequence are obtained; based on the data distribution characteristics of the suspected abnormal data points, similar data points are obtained; based on the quantity characteristics of similar data points, the data difference characteristics of similar data points, and the time interval characteristics, the degree of significant fluctuation of suspected abnormal data points is obtained. The degree of significant anomaly of the suspected abnormal data points is obtained based on the differences in the degree of significant fluctuations at the same position in the suspected abnormal data points and other electricity consumption behavior characteristic sequences, as well as the differences in local data changes; the user's billing risk coefficient is obtained based on the number of suspected abnormal data points, the degree of significant anomaly, and the time interval from the current date in the two electricity consumption behavior characteristic sequences. The electricity bill monitoring range is adaptively adjusted based on the aforementioned accounting risk coefficient to obtain the user's adaptive monitoring range; accounting reconciliation and monitoring are then performed based on the aforementioned adaptive monitoring range.
2. The intelligent verification and supervision method for marketing accounts of power enterprises according to claim 1, characterized in that, The step of obtaining the degree of suspected abnormal fluctuation of data points based on the data discrete characteristics and local data fluctuation characteristics of the electricity consumption behavior characteristic sequence includes: In the formula This indicates the degree of suspected abnormal fluctuations at the i-th data point. Indicates normalization, Let represent the value of the i-th data point, and H represent the average value of the electricity consumption behavior characteristic sequence. Indicates the first The value of each data point. express The value of each data point.
3. The intelligent verification and supervision method for marketing accounts of power enterprises according to claim 1, characterized in that, The step of obtaining suspected abnormal data points in the electricity consumption behavior feature sequence based on the suspected abnormal fluctuation degree includes: Data points in the electricity consumption behavior feature sequence whose suspected abnormal fluctuations exceed a preset threshold are considered as suspected abnormal data points.
4. The intelligent verification and supervision method for marketing accounts of power enterprises according to claim 1, characterized in that, The step of obtaining similar data points based on the data distribution characteristics of the suspected abnormal data points includes: Based on the values of the suspected abnormal data points, the suspected abnormal data points are clustered to obtain different clusters; suspected abnormal data points within the same cluster are considered as similar data points.
5. The intelligent verification and supervision method for marketing accounts of power enterprises according to claim 1, characterized in that, The step of obtaining the significant fluctuation degree of the suspected abnormal data points based on the quantity characteristics of similar data points, the data difference characteristics of similar data points, and the time interval characteristics of the suspected abnormal data points includes: In the formula This indicates the degree of significant fluctuation of the p-th suspected outlier data point. Let N represent an exponential function with the natural constant as the base, and let N represent the number of similar data points to the suspected outlier data point. Let F represent the value of the nth similar data point, and let F represent the average value of the similar data points of the suspected abnormal data point. This indicates the number of intervals between adjacent similar data points. G represents the length of the interval between the m-th adjacent data points, and G represents the average interval length between adjacent similar data points.
6. The intelligent verification and supervision method for marketing accounts of power enterprises according to claim 1, characterized in that, The step of obtaining the significant anomaly degree of the suspected abnormal data point based on the difference characteristics of significant fluctuations at the same position in the suspected abnormal data point and other electricity consumption behavior characteristic sequences, and the difference characteristics of local data changes, includes: In the formula This indicates the degree of significant anomaly of the p-th suspected outlier data point. This indicates the degree of significant fluctuation in suspected outlier data points. This indicates the degree of significant fluctuation of suspected abnormal data points at the same position in other electricity consumption behavior characteristic sequences. This represents the slope of the change between the p-th suspected outlier data point and the previous data point. This represents the slope of the change from the previous data point at the same position in other electricity consumption behavior characteristic sequences.
7. The intelligent verification and supervision method for marketing accounts of power enterprises according to claim 1, characterized in that, The step of obtaining the user's billing risk coefficient based on the quantity characteristics, degree of significant anomaly, and time interval characteristics of suspected abnormal data points in two electricity consumption behavior characteristic sequences includes: In the formula, E represents the user's billing risk coefficient, Q represents the number of electricity consumption behavior feature sequences, and U represents the number of suspected abnormal data points in the electricity consumption behavior feature sequences. This indicates the degree of significant anomaly of the p-th suspected outlier data point. This represents the time interval between the p-th suspected abnormal data point and the current date.
8. The intelligent verification and supervision method for marketing accounts of power enterprises according to claim 1, characterized in that, The step of adaptively adjusting the electricity bill monitoring interval based on the accounting risk coefficient to obtain the user's adaptive monitoring interval includes: , In the formula This indicates the adaptive minimum electricity cost. Indicates the adaptive maximum electricity price, 'a' represents the preset minimum electricity price, 'b' represents the preset maximum electricity price, and 'E' represents the accounting risk coefficient. and Construct adaptive monitoring intervals for users.