Electric energy meter user electricity consumption behavior prediction method

Abstract

The invention relates to the technical field of data processing, in particular to an electric energy meter user electricity consumption behavior prediction method, which comprises the following steps of: firstly, judging whether the electricity consumption behavior performance of a target user is sufficient or not, and if the electricity consumption behavior performance is insufficient, determining whether the electricity consumption behavior performance is sufficient in historical data of other users; acquiring a reference historical data sequence similar to the power consumption condition of the target user in the current time period; classifying all the users according to each user feature, and obtaining the association degree of each user feature and the user power consumption behavior according to the power consumption condition similarity of the users in each category of each user feature; and obtaining a target category of other users corresponding to each reference historical data sequence and the target user, and predicting load data of the target user in a future time period according to the association degree corresponding to the user feature to which each target category belongs and data in a time period after the time period of each reference historical data sequence. The accuracy of predicting the electricity consumption behavior of the electric energy meter user is improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method for predicting electricity consumption behavior of electricity meter users. Background Art

[0002] Electricity meter user behavior prediction involves analyzing historical user data, usage habits, and influencing factors to predict future usage. This prediction can be used to optimize power resource scheduling and identify abnormal usage, making it crucial for intelligent grid management.

[0003] Traditional methods for predicting electricity meter user behavior primarily train prediction models based on the user's historical usage data. The model then predicts the user's behavior based on the resulting prediction model and current data. However, when a new user or a user's behavior changes, the lack of sufficient historical data with similar usage patterns often leads to inaccurate predictions. This can lead to misidentification of normal usage as abnormal, impacting the effectiveness of power system management.

[0004] Therefore, how to improve the accuracy of predicting the electricity consumption behavior of electricity meter users has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, an embodiment of the present invention provides a method for predicting the electricity usage behavior of an electricity meter user, so as to solve the problem of how to improve the accuracy of predicting the electricity usage behavior of an electricity meter user.

[0006] An embodiment of the present invention provides a method for predicting electricity consumption behavior of an electric energy meter user, the method comprising the following steps: Based on the amount of power load data of the target user before the current period and the data change pattern, determine whether the target user's power consumption behavior is sufficient; If the power consumption behavior performance is insufficient, obtaining at least one reference historical data sequence based on the similarity between the power load data of at least one other user within a preset time range before the current time period and the power load data of the target user within the current time period; Obtain at least one user feature used to form a user profile, and for any user feature, classify the target user and all other users into at least one category based on the user feature, and obtain the degree of correlation between the user feature and the user's electricity usage behavior based on the similarity of changes in power load data of each user in each category before the current time period; The target categories of other users and the target user corresponding to each reference historical data sequence are obtained respectively. According to the degree of correlation between the user characteristics of each target category and the power load data in a period after the period of each reference historical data sequence, the predicted load data of the target user at each moment in the future period is obtained.

[0007] Preferably, judging whether the target user's electricity consumption behavior is sufficient based on the amount of power load data of the target user before the current period and the data change pattern includes: If the amount of power load data of the target user before the current period is less than a preset amount, it is determined that the target user's power consumption behavior performance is insufficient; If the number of power load data of the target user before the current period is greater than or equal to a preset number, the preset number of power load data of the target user before the current period are decomposed using the STL algorithm to obtain seasonal terms and trend terms, and curve fitting is performed on the seasonal terms and trend terms respectively to obtain a seasonal fitting curve and a trend fitting curve; Obtaining seasonal data and trend data of the target user at each moment in the current period according to the seasonal fitting curve and the trend fitting curve, and adding the seasonal data and trend data of the target user at each moment in the current period to obtain a load forecast value of the target user at each moment in the current period; Obtain the actual load value of the target user at each moment in the current period, calculate the absolute value of the difference between the load forecast value and the actual load value of the target user at each moment in the current period, obtain the prediction error, and accumulate all prediction errors to obtain the deviation of the target user's power consumption pattern in the current period; The actual load values ​​of the target user at each moment in the current time period are combined into a current load sequence, the absolute value of the difference between each two adjacent data in the current load sequence is calculated, all the absolute values ​​of the difference are accumulated to obtain the degree of data anomaly, the reciprocal of the sum of a preset constant and the degree of data anomaly is calculated, and the product of the reciprocal and the degree of deviation from the power consumption pattern is linearly normalized to obtain the degree of change in the power consumption behavior of the target user; If the degree of change in the electricity usage behavior is greater than or equal to a preset electricity usage behavior change degree threshold, it is determined that the target user's electricity usage behavior performance is insufficient.

[0008] Preferably, obtaining at least one reference historical data sequence based on similarity between power load data of at least one other user within a preset time range before the current time period and power load data of the target user within the current time period includes: The power load data of each other user within a preset time range before the current period are respectively grouped into a first historical sequence, and the noise level of each data in each first historical sequence is obtained according to the data change difference in each first historical sequence; According to the time length of the current period, the first historical sequence corresponding to each other user is divided into at least one first historical subsequence, and the power load data of the target user at each moment in the current period is combined into a current load sequence; For any first historical subsequence, obtaining a similarity index between the first historical subsequence and the current load sequence based on the noise level of each data in the first historical subsequence and the data similarity between the first historical subsequence and the current load sequence; The similarity indices corresponding to all first historical subsequences are obtained, and the first historical subsequences corresponding to all similarity indices greater than a preset similarity index threshold are recorded as reference historical data sequences.

[0009] Preferably, obtaining the noise level of each data in each first historical sequence according to the data change difference in each first historical sequence includes: For any data in any first historical sequence, obtain neighboring data at two moments closest to the moment corresponding to the any data; The absolute values ​​of the differences between any data and the neighborhood data at the two moments are calculated respectively to obtain a first degree of difference and a second degree of difference, and the sum of the first degree of difference and the second degree of difference is linearly normalized to obtain the noise degree of any data.

[0010] Preferably, obtaining a similarity index between any first historical subsequence and the current load sequence based on the noise level of each data in any first historical subsequence and the data similarity between any first historical subsequence and the current load sequence includes: For any data in any first historical subsequence, according to the position of the any data in the any first historical subsequence, obtaining target data having the same position as the any data in the current load sequence, calculating an absolute value of a difference between the any data and the target data, and calculating the inverse of a sum of the absolute value of the difference and a preset constant to obtain a degree of similarity between the any data and the target data; Accumulating the noise levels of all data in any first historical subsequence to obtain an overall noise level of any first historical subsequence, calculating a ratio of the noise level of any data to the overall noise level, and subtracting the ratio from a constant 1 to obtain a similarity credibility between any data and the target data; Using the similarity credibility between the any data and the target data as the weight of the similarity degree, weighting the similarity degree to obtain the weighted similarity between the any data and the target data; The weighted similarities corresponding to all data in any first historical subsequence are calculated, and the accumulated values ​​of all weighted similarities are linearly normalized to obtain a similarity index between any first historical subsequence and the current load sequence.

[0011] Preferably, obtaining the degree of association between any user feature and the user's electricity usage behavior based on the similarity of changes in power load data of each user in each category before the current period includes: For any category, denoising the power load data of each user in the category within a preset time range before a current period is performed to obtain a second historical sequence of each user in the category, and dividing each second historical sequence into at least one second historical subsequence based on the length of the current period; Calculating the mean of all data in each of the second historical subsequences respectively, performing linear normalization on the mean values ​​corresponding to all the second historical subsequences, and obtaining the load mean value for the period of each of the second historical subsequences; Performing linear fitting on the data in each of the second historical subsequences to obtain a fitted straight line for each of the second historical subsequences, calculating the slope of the fitted straight line for each of the second historical subsequences, and linearly normalizing all the slopes to obtain a load trend value for the time period of each of the second historical subsequences; Constructing a scatter plot based on the load mean and load trend value of all second historical subsequences under any category, where the abscissa of the scatter plot represents the load mean and the ordinate represents the load trend value; calculating the Euclidean distance between every two data points in the scatter plot, summing the reciprocals of all Euclidean distances to obtain a cumulative value, and linearly normalizing the cumulative value to obtain a power usage similarity index for all users under any category; The average value of the electricity usage similarity index of all categories under the any user feature is calculated to obtain the degree of association between the any user feature and the user's electricity usage behavior.

[0012] Preferably, the step of respectively obtaining the target categories of the target user and other users corresponding to each reference historical data sequence, and obtaining the predicted load data of the target user at each moment in the future period based on the degree of association corresponding to the user characteristics of each target category and the power load data in a period after the period in which each reference historical data sequence is located, includes: For any reference historical data sequence, if any category contains both other users corresponding to the reference historical data sequence and the target user, then the category is recorded as the target category, all target categories are obtained, and the user features belonging to all target categories are recorded as target user features; Set the similarity value between the target user and other users corresponding to any reference historical data sequence under each target user feature to be a constant of 1, and set the similarity value between the target user and other users corresponding to any reference historical data sequence under each non-target user feature to be a constant of 0; The similarity value corresponding to each user feature is used as the weight of the corresponding correlation degree of each user feature, and the weighted average of the corresponding correlation degrees of all user features is calculated to obtain the electricity consumption similarity between other users corresponding to any reference historical data sequence and the target user; Obtaining the degree to which any reference historical data sequence reflects the target user's electricity usage behavior in the current time period based on environmental differences between the time period corresponding to any reference historical data sequence and the current time period, and the similarity between electricity usage of other users corresponding to any reference historical data sequence and the target user; The power load data in a period after the period in which each reference historical data sequence is located are respectively combined into a target historical sequence. According to the data in each target historical sequence and the reflection degree of the reference historical data sequence corresponding to each target historical sequence, the predicted load data of the target user at each moment in the future period is obtained.

[0013] Preferably, obtaining the degree to which any reference historical data sequence reflects the target user's electricity usage behavior in the current time period based on the environmental differences between the time period corresponding to any reference historical data sequence and the current time period, and the electricity usage similarity between other users corresponding to any reference historical data sequence and the target user, includes: Obtain the historical temperature average and the historical humidity average within the corresponding time period of any reference historical data sequence, as well as the real-time temperature average and the real-time humidity average within the current time period; Calculating the absolute value of the difference between the historical temperature mean and the real-time temperature mean to obtain a temperature difference, calculating the reciprocal of a preset constant and the temperature difference to obtain temperature similarity, obtaining humidity similarity based on the historical humidity mean and the real-time humidity mean, and linearly normalizing the sum of the temperature similarity and the humidity similarity to obtain environmental similarity between a corresponding time period of any reference historical data sequence and the current time period; The product of the electricity usage similarity between other users and the target user corresponding to any reference historical data sequence and the environmental similarity is calculated to obtain the degree to which any reference historical data sequence reflects the electricity usage behavior of the target user in the current time period.

[0014] Preferably, obtaining the predicted load data of the target user at each moment in the future period based on the data in each target historical sequence and the reflection degree of the reference historical data sequence corresponding to each target historical sequence includes: Calculate the cumulative value of the reflection degree of the reference historical data series corresponding to all target historical series to obtain a comprehensive reflection degree, and calculate the proportion of the reflection degree of the reference historical data series corresponding to each target historical series in the comprehensive reflection degree to obtain a reflection weight of each target historical series; Each moment of the target user in the future time period is organized into a time series. For any moment in the time series, according to the position of any moment in the time series, target historical data with the same position as that of any moment is obtained in each target historical sequence. Each target historical data is weighted according to the reflection weight of the target historical sequence in which each target historical data is located to obtain a weighted average value of all target historical data. The weighted average value is used as the predicted load data of the target user at any moment in the future time period.

[0015] Preferably, the user characteristics include the user's electricity usage type, the user's house area, the user's floor, and the user's region.

[0016] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: Based on the amount of power load data of the target user before the current time period and the data change pattern, determine whether the power consumption behavior performance of the target user is sufficient; if the power consumption behavior performance is insufficient, obtain at least one reference historical data sequence based on the similarity between the power load data of at least one other user within a preset time range before the current time period and the power load data of the target user in the current time period; obtain at least one user feature for forming a user portrait, and for any user feature, divide the target user and all other users into at least one category based on the any user feature, and obtain the degree of correlation between the any user feature and the user's power consumption behavior based on the similarity of the change of power load data of each user in each category before the current time period; obtain the target category of the other users and the target user corresponding to each reference historical data sequence respectively, and obtain the predicted load data of the target user at each moment in the future time period based on the corresponding degree of correlation of the user features belonging to each target category and the power load data in a time period after the time period of each reference historical data sequence. Among them, when the target user's electricity consumption behavior is insufficient, a reference historical data sequence with similar load data change characteristics to the target user in the current period is obtained from the historical load data of other users, and then the load data of the target user in the future period is predicted based on the data of the next period of the period where each reference historical data sequence is located, so as to effectively solve the problem of inaccurate prediction results caused by insufficient user historical electricity consumption data or changes in user electricity consumption behavior; in the process of predicting the load data of the target user in the future period based on the data of the next period of the period where each reference historical data sequence is located, the target user and all other users are classified according to user characteristics, so as to find other users with similar user characteristics to the target user among the other users corresponding to each reference historical data sequence, and at the same time, according to the similarity of the load data of users under each category, the degree of correlation between each user characteristic and electricity consumption behavior is obtained, so as to adjust the influence of each reference historical data sequence on the prediction result, reduce the interference of similar historical data caused by accidental factors on the prediction result, and improve the accuracy of predicting the electricity consumption behavior of electricity meter users. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 This is a flow chart of a method for predicting electricity consumption behavior of an electric energy meter user provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0019] The embodiments of the present disclosure are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to be used to explain the present disclosure, but should not be understood as limiting the present disclosure.

[0020] It should be noted that the terms "first," "second," and the like in the specification of the present disclosure and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure.

[0021] In order to illustrate the technical solution of the present invention, specific embodiments are provided below.

[0022] See also Figure 1 , is a flow chart of a method for predicting the electricity consumption behavior of an electric energy meter user provided in the first embodiment of the present invention, such as Figure 1 As shown, the method may include: Step S101 : judging whether the target user's electricity consumption behavior is sufficient based on the amount of the target user's power load data before the current period and the data variation pattern.

[0023] Any user whose electricity usage behavior is to be predicted is designated as a target user. Traditional methods for predicting electricity meter user electricity usage behavior primarily train a prediction model based on the target user's historical electricity usage data. Based on the obtained prediction model, the target user's electricity usage data for the current period is used to predict the target user's electricity usage behavior. However, when the target user is new or their electricity usage behavior has changed, there is often insufficient historical data with similar electricity usage patterns, resulting in insufficient representation of the target user's electricity usage behavior, which in turn leads to inaccurate prediction results.

[0024] Therefore, it is first necessary to determine whether the target user's electricity consumption behavior is sufficient, that is, to determine whether the target user is a new user or a user whose electricity consumption behavior has changed. If the target user's electricity consumption behavior is sufficient, that is, the target user is not a new user and the electricity consumption behavior has not changed, then it means that the target user's historical electricity consumption is sufficient to reflect the target user's electricity consumption behavior. At this time, the target user's electricity consumption behavior can be predicted in a traditional way, that is, the prediction model is trained based on the target user's historical electricity consumption data, and the target user's electricity consumption data in the current time period is used to predict the target user's electricity consumption behavior in the future time period based on the obtained prediction model. The time period is set to 1 hour, that is, the target user's electricity consumption behavior in the next 1 hour is predicted based on the obtained prediction model and the target user's data in the current 1 hour. The prediction model can be based on LSTM, XGBoost, ARIMA, etc., and there is no restriction here. The implementer can set it according to the specific scenario. The training of the prediction model is an existing technology and will not be repeated here.

[0025] If the target user's electricity usage behavior is not sufficiently represented, that is, the target user is a new user or a user whose electricity usage behavior has changed, then in an embodiment of the present invention, the target user's electricity usage behavior in the future time period is predicted based on other users with similar electricity usage behaviors as the target user under the same power supply system.

[0026] The process of judging whether the target user's electricity consumption behavior is sufficient is as follows: First, determine whether the target user is a new user: In an embodiment of the present invention, the user's load data within 7 days before the current time period is used as historical electricity consumption data for analysis, wherein the load data is collected once every minute. There is no restriction here. The implementer can set the time range and data collection frequency of the historical electricity consumption data according to the specific scenario. If the target user's electricity load data before the current time period is less than 7 days, that is, the number of data of the target user's electricity load data before the current time period is less than 10080 (that is, the preset number, 7×24×60=10080), it is determined that the target user's electricity consumption behavior is insufficient; it should be noted that, considering that the general residents work, go to school, and factories operate in a cycle of one week, 7 days can fully reflect the user's electricity consumption pattern, and the electricity load data is collected once a minute, the electricity load data collected for 7 days should be 10080. Therefore, when the number of historical electricity load data for a user is less than 7 days, it will be defaulted to a new user.

[0027] If the number of data of the target user's power load data before the current period is greater than or equal to 10080, it means that the target user is not a new user, but it is necessary to continue to determine whether the target user's electricity consumption behavior has changed: use the STL algorithm to decompose the target user's power load data within 7 days before the current period to obtain seasonal items and trend items. The seasonal cycle length in the STL algorithm is set to 1 day, that is, 1440 data. There is no restriction here. The implementer can use the least squares method to perform curve fitting on the seasonal item and trend item according to the specific scenario settings to obtain a seasonal fitting curve and a trend fitting curve. The STL algorithm and the least squares method are existing technologies and will not be described in detail here. According to the seasonal fitting curve and the trend fitting curve, the seasonal data and trend data of the target user at each moment in the current period are obtained. The seasonal data and trend data of the target user at each moment in the current period are added together to obtain the load forecast value of the target user at each moment in the current period. It is worth noting that the residual decomposed by the STL algorithm is regarded as noise data and is therefore not involved in the prediction. Obtain the actual load value of the target user at each moment (i.e., every minute) in the current time period, calculate the absolute value of the difference between the load forecast value and the actual load value of the target user at each moment in the current time period, obtain the prediction error, accumulate all prediction errors, and obtain the deviation of the target user's electricity consumption pattern in the current time period, which is used to determine whether the target user's electricity consumption behavior has changed.

[0028] In one embodiment, the calculation formula for the target user's electricity usage pattern deviation in the current time period is:

[0029] Wherein, Q represents the deviation of the target user's power consumption pattern in the current period, N represents the number of actual load values ​​in the current period (in the embodiment of the present invention, N=60), represents the load forecast value at the i-th moment in the current period, represents the actual load value at the i-th moment in the current period, Indicates the absolute value symbol.

[0030] It should be noted that The larger Q is, the greater the difference between the target user's electricity usage pattern in the current period and the historical electricity usage pattern before the current period, and the less the target user's electricity usage pattern in the current period conforms to the target user's historical electricity usage pattern. Therefore, the larger Q is, the more likely it is that the target user's electricity usage behavior in the current period has changed.

[0031] Considering that the load data mutation caused by the electric energy meter failure will also cause a large difference between the target user's electricity consumption pattern in the current period and the historical electricity consumption pattern, and the load data mutation caused by the electric energy meter failure is a data anomaly, not a change in the user's electricity consumption behavior, so relying solely on the electricity consumption pattern deviation Q cannot accurately determine whether the target user's electricity consumption behavior has changed. The degree of data anomaly of the target user's actual load value in the current period should also be combined to obtain the degree of change in the target user's electricity consumption behavior, and thus more accurately determine whether the target user's electricity consumption behavior has changed. Specifically: The actual load values ​​of the target user at each moment in the current time period are combined into a current load sequence, the absolute value of the difference between each two adjacent data in the current load sequence is calculated, all the absolute values ​​of the difference are accumulated to obtain the degree of data anomaly, the reciprocal of the sum of the preset constant and the degree of data anomaly is calculated, the product of the reciprocal and the degree of deviation from the power consumption pattern is linearly normalized to obtain the degree of change in the power consumption behavior of the target user, wherein linear normalization is an existing technology and will not be repeated here.

[0032] In one embodiment, the calculation formula for the degree of change in the target user's electricity usage behavior is:

[0033] Where P represents the degree of change in the target user's electricity consumption behavior, Q represents the deviation of the target user's electricity consumption pattern in the current period, and N represents the number of actual load values ​​in the current period (in this embodiment of the present invention, N=60). Indicates the actual load value at the i-th moment in the current period, that is, the i-th actual load value in the current load sequence. Indicates the actual load value at the i-1th moment in the current period, that is, the i-1th actual load value in the current load sequence. Represents a preset constant, used to prevent the denominator from being 0. In the embodiment of the present invention, set ,There is no restriction here, and the implementer can set it according to the specific scenario. Indicates the absolute value symbol, represents the linear normalization function.

[0034] It should be noted that The smaller it is, the larger Q is, which means that the possibility of abnormal data caused by electricity meter failure in the current load sequence is smaller, the greater the difference between the target user's electricity consumption pattern in the current period and the historical electricity consumption pattern before the current period, the greater the authenticity of the change in the target user's electricity consumption behavior in the current period, and thus the larger P is, the more likely the target user's electricity consumption behavior in the current period has changed.

[0035] Whether the user's electricity consumption behavior has changed varies greatly in the electricity consumption data. After normalization, the two situations of the degree of change in electricity consumption behavior are divided on both sides of the [0, 1] interval. Therefore, the middle value is taken as the division threshold, that is, the preset threshold of the degree of change in electricity consumption behavior is set to 0.5. There is no restriction here. The implementer can set it according to the specific scenario. If P ≥ 0.5, it is determined that the target user's electricity consumption behavior has changed in the current period, and the target user's electricity consumption behavior performance is insufficient.

[0036] At this point, the judgment on whether the target user's electricity consumption behavior is sufficient is completed.

[0037] Step S102: If the electricity consumption behavior is insufficient, at least one reference historical data sequence is obtained based on the similarity between the power load data of at least one other user within a preset time range before the current period and the power load data of the target user in the current period.

[0038] If the target user's electricity usage behavior is not sufficiently demonstrated, that is, the target user is a new user or a user whose electricity usage behavior has changed, then in an embodiment of the present invention, at least one reference historical data sequence similar to the target user's electricity usage behavior in the current time period is obtained from the power load data of each other user under the power supply system where the target user is located within 7 days before the current time period (there is no restriction here, and the implementer can set the time range according to the specific scenario), and the target user's electricity usage behavior in the future time period is subsequently predicted based on the load data in the time period following the time period where each reference historical data sequence is located.

[0039] Since the power load data of each other user in the 7 days before the current period is likely to contain noise data, this noise data will affect the similarity judgment. Therefore, first, the power load data of each other user in the 7 days before the current period are respectively formed into a first historical sequence. Based on the data change differences in each first historical sequence, the noise level of each data in each first historical sequence is obtained to reduce the noise impact in the subsequent data analysis process. Taking the jth data in the first historical sequence of the uth other user as an example, the specific method for obtaining the noise level of the jth data in the first historical sequence of the uth other user is: In the first historical sequence of the u-th other user, obtain the neighborhood data of the two moments closest to the moment corresponding to the j-th data; The absolute values ​​of the differences between the j-th data and the neighborhood data at the two moments are calculated respectively to obtain a first degree of difference and a second degree of difference. The sum of the first degree of difference and the second degree of difference is linearly normalized to obtain the noise level of the j-th data in the first historical sequence of the u-th other user.

[0040] In one embodiment, the calculation formula for the noise level of the jth data in the first historical sequence of the uth other user is:

[0041] in, represents the noise level of the jth data in the first historical sequence of the uth other user, Represents the first neighboring data in the two nearest neighboring data moments corresponding to the jth data moment, Represents the second neighboring data in the two nearest neighboring data corresponding to the jth data. represents the jth data in the first historical sequence of the uth other user, Indicates the absolute value symbol, represents the linear normalization function.

[0042] It should be noted that The larger the value, the greater the mutation degree of the jth data. The larger , the more likely the jth data is noise data.

[0043] Similarly, the noise level of each data in the first historical sequence of each other user is obtained.

[0044] Furthermore, according to the duration of the current time period, that is, 1 hour, the first historical sequence corresponding to each other user is divided into at least one first historical subsequence (in an embodiment of the present invention, the first historical sequence corresponding to each other user is the power load data within 7 days before the current hour, so the number of first historical subsequences corresponding to each other user is 168). Based on the current load sequence of the target user in the current time period, the similarity between each first historical subsequence and the current load sequence is analyzed to obtain a similarity index between each first historical subsequence and the current load sequence, and then find at least one reference historical data sequence that is similar to the target user's electricity usage behavior in the current time period.

[0045] Taking the vth first historical subsequence of the uth other user as an example, based on the noise level of each data in the vth first historical subsequence of the uth other user and the data similarity between the vth first historical subsequence of the uth other user and the current load sequence, the similarity index between the vth first historical subsequence of the uth other user and the current load sequence is obtained. Specifically: For any data in the vth first historical subsequence of the uth other user, based on the position of the any data in the vth first historical subsequence of the uth other user, obtain target data in the current load sequence that has the same position as the any data, calculate the absolute value of the difference between the any data and the target data, calculate the inverse of the sum of the absolute value of the difference and a preset constant, and obtain the degree of similarity between the any data and the target data; Accumulate the noise levels of all data in the vth first historical subsequence of the uth other user to obtain the overall noise level of the vth first historical subsequence of the uth other user, calculate the proportion of the noise level of any data in the overall noise level, and subtract the proportion from a constant 1 to obtain the similarity credibility between the any data and the target data; Using the similarity credibility between the any data and the target data as the weight of the similarity degree, weighting the similarity degree to obtain the weighted similarity between the any data and the target data; The weighted similarities corresponding to all data in the vth first historical subsequence of the uth other user are calculated, and the accumulated values ​​of all weighted similarities are linearly normalized to obtain a similarity index between the vth first historical subsequence of the uth other user and the current load sequence.

[0046] In one embodiment, the calculation formula for the similarity index between the vth first historical subsequence of the uth other user and the current load sequence is:

[0047] in, represents the similarity index between the vth first historical subsequence of the uth other user and the current load sequence, N represents the number of data in the current load sequence (in this embodiment of the present invention, N=60), and the number of data in the vth first historical subsequence of the uth other user is equal to the number of data in the current load sequence. represents the zth data in the vth first historical subsequence of the uth other user, Indicates the zth data in the current load sequence, Represents a preset constant, used to prevent the denominator from being 0. In the embodiment of the present invention, set ,There is no restriction here, and the implementer can set it according to the specific scenario. represents the noise level of the jth data in the vth first historical subsequence of the uth other user, represents the noise level of the zth data in the vth first historical subsequence of the uth other user, Indicates the absolute value symbol, represents the linear normalization function.

[0048] It should be noted that The smaller is , the greater the similarity between the vth first historical subsequence of the uth other user and the current load sequence. The smaller the value, the more credible the similarity between the vth first historical subsequence of the uth other user and the current load sequence is. The larger the value is, the more similar the vth first historical subsequence of the uth other user is to the current load sequence, and the more similar the electricity consumption behavior of the uth other user in the period of the vth first historical subsequence is to the electricity consumption behavior of the target user in the current period.

[0049] Similarly, the similarity index between each first historical subsequence of each other user and the current load sequence is obtained. According to the distribution of abnormal data in the experimental data, it is found that abnormal data and normal data are usually distributed on both sides of 0.7. Therefore, the preset similarity index threshold is set to 0.7. There is no restriction here. The implementer can obtain the first historical subsequences corresponding to all similarity indicators greater than 0.7 according to the specific scenario settings, and record them as reference historical data sequences similar to the electricity consumption behavior of the target user in the current time period. They are used to predict the electricity consumption behavior of the target user in the future time period based on the load data in the next time period after the time period of each reference historical data sequence.

[0050] Step S103: obtain at least one user feature for forming a user portrait, and for any user feature, divide the target user and all other users into at least one category based on the any user feature, and obtain the degree of correlation between the any user feature and the user's electricity usage behavior based on the similarity of changes in the power load data of each user in each category before the current time period.

[0051] Through step S102, a reference historical data sequence similar to the target user's electricity usage behavior in the current time period is obtained. However, since the reference historical data sequence represents the electricity usage of other users in a certain historical time period, even if the electricity usage of other users in a certain historical time period is similar to the target user's electricity usage in the current time period, their electricity usage in the future is likely to be different. Considering that users with different user characteristics may have different electricity usage behaviors, for example, in apartment-type housing, the first three floors of an apartment-type housing are generally used for commercial electricity, while the fourth floor and above are used for residential electricity. Residential electricity consumption will have peaks in the morning and evening, while commercial electricity consumption may be stable throughout the day. For another example, in industrial production workshops, large industrial production workshops generally have more production equipment than small industrial production workshops, and large industrial production workshops also use more electricity than small industrial production workshops.

[0052] Therefore, in an embodiment of the present invention, first, the user's electricity usage type (such as household electricity usage, general commercial electricity usage, agricultural production electricity usage, etc.), the user's house area (such as small apartment, medium apartment, large apartment, etc.), the floor where the user is located (such as low floor, middle floor, high floor, etc.), and the area where the user is located are used to form the user features of the user portrait, and all users under the power supply system where the target user is located are divided into at least one category based on each user feature. If any two users belong to the same category, it means that the two users have similar user features. Among them, all users under the power supply system where the target user is located are divided into at least one category based on each user feature. For example: based on the user's electricity usage type, all users are divided into household electricity users, general commercial electricity users, agricultural production electricity users, etc.; based on the user's house area, all users are divided into small apartment users (for example, less than 60 square meters are small apartments), medium apartment users (for example, 60 to 120 square meters are medium apartments), large apartment users (for example, 120 to 200 square meters are large apartments), etc.; based on the user's floor, all users are divided into low-level users (for example, 1 to 3rd floor for low-level users), middle-level users (for example, floors 4 to 12 for middle-level users), high-level users (for example, floors 13 and above for high-level users), etc.; based on the user's region, the longitude and latitude of all user locations are mapped into a scatter plot, with the horizontal axis of the scatter plot representing longitude and the vertical axis representing latitude. The DBSCAN clustering algorithm is used to cluster the data points in the scatter plot to obtain at least one cluster. Users belonging to the same cluster are considered to be users of the same region. Users whose houses face the same direction can also be considered to be users of the same region, or all users can be classified according to official regional divisions. There are no restrictions here. Implementers can set the number of user features, as well as the classification method and classification criteria for each user feature, based on the specific scenario. The DBSCAN clustering algorithm is an existing technology and will not be described in detail here.

[0053] Then, based on the similarity of changes in power load data of users belonging to the same category under different user characteristics before the current period, the degree of correlation between each user characteristic and the user's electricity consumption behavior is obtained. If the correlation degree of a certain user characteristic is large, it means that users with similar characteristics under this user characteristic also have similar electricity consumption behaviors. That is, if other users corresponding to any reference historical data sequence belong to the same category as the target user under any user characteristic, and the correlation degree of any user characteristic is large, it means that other users corresponding to any reference historical data sequence have similar electricity consumption behaviors as the target user. It can be considered that the electricity consumption behavior in the time period of any reference historical data sequence is not only similar to the electricity consumption behavior of the target user in the current period, but also the electricity consumption behavior in a period after the corresponding period of any reference historical data sequence is also likely to be similar to the electricity consumption behavior of the target user in the future period. Therefore, when predicting the target user's electricity consumption behavior in the future period based on the data of the period after the period of any reference historical data sequence, it should be given a higher weight to adjust the influence of any reference historical data sequence on the prediction result, thereby improving the accuracy of predicting the target user's electricity consumption behavior in the future period.

[0054] To obtain the degree of correlation between each user feature and the user's electricity usage behavior, take the ath user feature as an example. Based on the similarity of the change of power load data of all users under each category of the ath user feature before the current period, the degree of correlation between the ath user feature and the user's electricity usage behavior is obtained. Specifically: (1) According to the similarity of the change of power load data of users in each category of the a-th user feature before the current period, the power consumption similarity index of all users in each category of the a-th user feature is obtained.

[0055] Specifically: for the t-th category of the a-th user feature, use median filtering to perform denoising on the power load data of each user in the t-th category within 7 days before the current period (this is not limited here, and the implementer can set the time range according to the specific scenario), thereby obtaining a second historical sequence for each user in the t-th category. Based on the length of the current period (in this embodiment of the present invention, the length of the current period is 1 hour), each second historical sequence is divided into at least one second historical subsequence. The use of median filtering for denoising is a conventional technique and will not be described in detail here. Calculating the mean of all data in each of the second historical subsequences respectively, performing linear normalization on the mean values ​​corresponding to all the second historical subsequences, and obtaining the load mean value for the period of each of the second historical subsequences; Performing linear fitting on the data in each of the second historical subsequences using the least squares method to obtain a fitted straight line for each of the second historical subsequences, calculating the slope of the fitted straight line for each of the second historical subsequences, and linearly normalizing all the slopes to obtain a load trend value for the time period of each of the second historical subsequences; A scatter plot is constructed based on the load mean and load trend value of all second historical subsequences under the t-th category, where the horizontal axis of the scatter plot represents the load mean and the vertical axis represents the load trend value. In the scatter plot, the Euclidean distance between every two data points is calculated, that is, the data in the scatter plot are grouped in pairs, the Euclidean distance between the two data in each group is calculated, the reciprocals of all Euclidean distances are accumulated to obtain an accumulated value, and the accumulated value is linearly normalized to obtain a similarity index of electricity usage for all users under the t-th category.

[0056] In one embodiment, the calculation formula for the electricity usage similarity index of all users in the t-th category is:

[0057] in, represents the similarity index of electricity consumption of all users in the t-th category, It represents the number of all data points in the scatter plot constructed based on the load mean and load trend values ​​of all second historical subsequences under the t-th category, that is, the number of all second historical subsequences under the t-th category. That is, the number of combinations of all the data in the scatter plot. represents the Euclidean distance between the two data in the cth combination, represents the linear normalization function.

[0058] It should be noted that The smaller it is, the denser the data distribution in the scatter plot is, and the more similar the load data characteristics of users in the t-th category are. The larger , the more similar the electricity usage of users in the tth category.

[0059] Similarly, the electricity usage similarity index of each category of the a-th user feature is obtained.

[0060] (2) According to the electricity usage similarity index of each category of the a-th user feature, the correlation degree between the a-th user feature and the user's electricity usage behavior is obtained.

[0061] Specifically, the average value of the electricity usage similarity indexes of all categories under the a-th user feature is calculated to obtain the correlation degree between the a-th user feature and the user's electricity usage behavior.

[0062] In one embodiment, the calculation formula for the correlation between the ath user feature and the user's electricity usage behavior is:

[0063] in, Indicates the correlation between the ath user feature and the user's electricity consumption behavior, represents the number of all categories under the a-th user feature, Represents the similarity index of electricity usage under the t-th category of the a-th user feature.

[0064] It should be noted that The larger the value is, the more similar the electricity consumption of users in each category of the a-th user feature is. The larger it is, the greater the correlation between the ath user feature and the user's electricity usage behavior.

[0065] Similarly, the degree of correlation between each user feature and the user's electricity usage behavior is obtained.

[0066] Step S104, respectively obtain the target categories of other users and the target user corresponding to each reference historical data sequence, and obtain the predicted load data of the target user at each moment in the future period based on the degree of correlation corresponding to the user characteristics of each target category and the power load data in a period after the period of each reference historical data sequence.

[0067] The degree of correlation between each user feature and the user's electricity usage behavior is obtained through step S103, and the reference historical data sequence obtained in step S102 is historical data of other users in a certain historical period whose electricity usage is similar to that of the target user in the current period. If the other users corresponding to any reference historical data sequence belong to the same category as the target user under a certain user feature, and the degree of correlation between the user feature and the user's electricity usage behavior is relatively high, it is considered that the electricity usage behavior in a period after the period corresponding to any reference historical data sequence is also likely to be similar to the electricity usage behavior of the target user in the future period. The greater the degree to which any reference historical data sequence reflects the electricity usage behavior of the target user, the higher the weight should be given when predicting the target user's electricity usage behavior in the future period based on the data of the period after the period in which any reference historical data sequence is located. At the same time, considering that changes in the environment such as temperature and humidity are also one of the important factors affecting users' electricity consumption, the closer the temperature and humidity of the time period of any reference historical data sequence are to the current time period, the greater the reflection of any reference historical data sequence on the target user's electricity consumption behavior. In the subsequent prediction of the target user's electricity consumption behavior in the future time period based on the data of the period after the time period of any reference historical data sequence, it should also be given a higher weight to improve the accuracy of the prediction of the target user's electricity consumption behavior in the future time period.

[0068] Regarding the degree to which the reference historical data sequence reflects the target user's electricity consumption behavior: Taking the vth reference historical data sequence as an example, if any category contains both other users corresponding to the vth reference historical data sequence and target users, then any category is recorded as the target category, all target categories are obtained, and the user features belonging to all target categories are recorded as target user features. That is, if the other users corresponding to the vth reference historical data sequence and the target user are classified into the same category under a certain user feature, then the user feature is recorded as the target user feature. According to the degree of correlation corresponding to each target user feature and the environmental difference between the time period corresponding to the vth reference historical data sequence and the current time period, the degree to which the vth reference historical data sequence reflects the target user's electricity consumption behavior in the current time period is obtained. The specific steps are as follows: (1) According to the degree of association corresponding to each target user’s characteristics, the electricity consumption similarity between other users and the target user corresponding to the vth reference historical data sequence is obtained.

[0069] Specifically: set the similarity value between the target user and other users corresponding to the vth reference historical data sequence under each target user feature to be a constant of 1, and set the similarity value between the target user and other users corresponding to the vth reference historical data sequence under each non-target user feature to be a constant of 0; The similarity value corresponding to each user feature is used as the weight of the corresponding association degree of each user feature, and the weighted mean of the corresponding association degrees of all user features is calculated to obtain the electricity consumption similarity between other users corresponding to the vth reference historical data sequence and the target user.

[0070] In one embodiment, the calculation formula for the electricity consumption similarity between the target user and other users corresponding to the vth reference historical data sequence is:

[0071] in, represents the electricity consumption similarity between other users and the target user corresponding to the vth reference historical data sequence, Indicates the number of user features. In the embodiment of the present invention, , respectively, the user's electricity type, the user's house area, the user's floor and the user's area, represents the similarity value (value is 1 or 0) between the target user and other users corresponding to the vth reference historical data sequence under the ath user feature, Indicates the correlation degree between the ath user feature and the user's electricity usage behavior.

[0072] It should be noted that The larger the value is, the more similar the electricity consumption behavior of other users corresponding to the vth reference historical data sequence is to the target user, and thus The larger it is, the greater the weight corresponding to the vth reference historical data sequence.

[0073] (2) Based on the electricity usage similarity between other users corresponding to the vth reference historical data sequence and the target user, and the environmental differences between the time period corresponding to the vth reference historical data sequence and the current time period, the degree to which the vth reference historical data sequence reflects the target user's electricity usage behavior in the current time period is obtained.

[0074] Specifically: using the meteorological platform of the area to which the power supply system belongs, according to the load data collection frequency in step S101, that is, once per minute, respectively obtain the temperature value and humidity value of the time period corresponding to the v-th reference historical data sequence and the current time period, and linearly normalize all the obtained temperature values ​​and humidity values ​​to obtain the corresponding temperature data and humidity data; based on the temperature data and humidity data of the time period corresponding to the v-th reference historical data sequence and the current time period, respectively calculate the historical temperature average and the historical humidity average in the time period corresponding to the v-th reference historical data sequence, as well as the real-time temperature average and the real-time humidity average in the current time period; Calculating the absolute value of the difference between the historical temperature mean and the real-time temperature mean to obtain a temperature difference, calculating the reciprocal of a preset constant and the temperature difference to obtain a temperature similarity, obtaining a humidity similarity based on the historical humidity mean and the real-time humidity mean, and linearly normalizing the sum of the temperature similarity and the humidity similarity to obtain an environmental similarity between the time period corresponding to the vth reference historical data sequence and the current time period; The product of the electricity usage similarity between other users and the target user corresponding to the vth reference historical data sequence and the environmental similarity is calculated to obtain the degree to which the vth reference historical data sequence reflects the electricity usage behavior of the target user in the current period.

[0075] In one embodiment, the calculation formula for the degree to which the vth reference historical data sequence reflects the target user's electricity usage behavior in the current period is:

[0076] in, It represents the degree to which the vth reference historical data sequence reflects the target user’s electricity consumption behavior in the current period. represents the electricity consumption similarity between other users and the target user corresponding to the vth reference historical data sequence, represents the historical temperature mean value in the corresponding period of the vth reference historical data sequence, Indicates the real-time average temperature in the current period. represents the historical humidity mean value in the corresponding period of the vth reference historical data sequence, Indicates the real-time average humidity value in the current period. Represents a preset constant, used to prevent the denominator from being 0. In the embodiment of the present invention, set ,There is no restriction here, and the implementer can set it according to the specific scenario. Indicates the absolute value symbol, represents the linear normalization function.

[0077] It should be noted that The smaller, The smaller, The larger the value is, the more similar the environment in the period corresponding to the vth reference historical data sequence is to the current period, and the more similar the electricity consumption behavior of other users corresponding to the vth reference historical data sequence is to the target user. The larger the value is, the more the vth reference historical data sequence reflects the target user's electricity consumption behavior in the current period, and the greater the weight corresponding to the vth reference historical data sequence.

[0078] At this point, the degree to which the vth reference historical data sequence reflects the target user's electricity usage behavior in the current period is obtained. Similarly, the degree to which each reference historical data sequence reflects the target user's electricity usage behavior in the current period is obtained.

[0079] Furthermore, based on the data of the period following the period in which each reference historical data sequence is located, the target user's electricity consumption behavior in the future period is predicted. First, the power load data in the period (1 hour) following the period in which each reference historical data sequence is located is combined into a target historical sequence. Then, based on the data in each target historical sequence and the reflection degree of the reference historical data sequence corresponding to each target historical sequence, the load data of the target user at each moment in the future period (i.e., the next 1 hour) is predicted to obtain the predicted load data of the target user at each moment in the future period. Specifically: Calculate the cumulative value of the reflection degree of the reference historical data series corresponding to all target historical series to obtain a comprehensive reflection degree, and calculate the proportion of the reflection degree of the reference historical data series corresponding to each target historical series in the comprehensive reflection degree to obtain a reflection weight of each target historical series; Each moment of the target user in the future time period is organized into a time series. For any moment in the time series, according to the position of any moment in the time series, target historical data with the same position as that of any moment is obtained in each target historical sequence. Each target historical data is weighted according to the reflection weight of the target historical sequence in which each target historical data is located to obtain a weighted average value of all target historical data. The weighted average value is used as the predicted load data of the target user at any moment in the future time period.

[0080] In one embodiment, taking the xth moment in the time series as an example, that is, taking the xth moment in the future period as an example, the calculation formula for the predicted load data of the target user at the xth moment in the future period is:

[0081] in, represents the forecast load data of the target user at the xth moment in the future period, m represents the number of reference historical data sequences, that is, the number of target historical sequences, represents the xth data in the target historical sequence corresponding to the vth reference historical data sequence, Indicates the reflection degree of the vth reference historical data sequence, Indicates the The degree of reflection of a reference historical data series.

[0082] Similarly, the predicted load data of the target user at each moment in the future period is obtained to complete the prediction of the target user's electricity consumption behavior in the future period.

[0083] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A method for predicting electricity consumption behavior of electric energy meter users, characterized in that: The method for predicting electricity consumption behavior of an electric energy meter user comprises: Based on the amount of power load data of the target user before the current period and the data change pattern, determine whether the target user's power consumption behavior is sufficient; If the power consumption behavior performance is insufficient, obtaining at least one reference historical data sequence based on the similarity between the power load data of at least one other user within a preset time range before the current time period and the power load data of the target user within the current time period; Obtain at least one user feature used to form a user profile, and for any user feature, classify the target user and all other users into at least one category based on the user feature, and obtain the degree of correlation between the user feature and the user's electricity usage behavior based on the similarity of changes in power load data of each user in each category before the current time period; The target categories of other users and the target user corresponding to each reference historical data sequence are obtained respectively. According to the degree of correlation between the user characteristics of each target category and the power load data in a period after the period of each reference historical data sequence, the predicted load data of the target user at each moment in the future period is obtained.

2. The method for predicting electricity consumption behavior of electric energy meter users according to claim 1, characterized in that: The method of judging whether the target user's electricity consumption behavior is sufficient based on the amount of the target user's power load data before the current period and the data change pattern includes: If the amount of power load data of the target user before the current period is less than a preset amount, it is determined that the target user's power consumption behavior performance is insufficient; If the number of power load data of the target user before the current period is greater than or equal to a preset number, the preset number of power load data of the target user before the current period are decomposed using the STL algorithm to obtain seasonal terms and trend terms, and curve fitting is performed on the seasonal terms and trend terms respectively to obtain a seasonal fitting curve and a trend fitting curve; Obtaining seasonal data and trend data of the target user at each moment in the current period according to the seasonal fitting curve and the trend fitting curve, and adding the seasonal data and trend data of the target user at each moment in the current period to obtain a load forecast value of the target user at each moment in the current period; Obtain the actual load value of the target user at each moment in the current period, calculate the absolute value of the difference between the load forecast value and the actual load value of the target user at each moment in the current period, obtain the prediction error, and accumulate all prediction errors to obtain the deviation of the target user's power consumption pattern in the current period; The actual load values ​​of the target user at each moment in the current time period are combined into a current load sequence, the absolute value of the difference between each two adjacent data in the current load sequence is calculated, all the absolute values ​​of the difference are accumulated to obtain the degree of data anomaly, the reciprocal of the sum of a preset constant and the degree of data anomaly is calculated, and the product of the reciprocal and the degree of deviation from the power consumption pattern is linearly normalized to obtain the degree of change in the power consumption behavior of the target user; If the degree of change in the electricity usage behavior is greater than or equal to a preset electricity usage behavior change degree threshold, it is determined that the target user's electricity usage behavior performance is insufficient.

3. The method for predicting electricity consumption behavior of electric energy meter users according to claim 1, characterized in that: The obtaining of at least one reference historical data sequence based on similarity between power load data of at least one other user within a preset time range before the current time period and power load data of the target user within the current time period includes: The power load data of each other user within a preset time range before the current period are respectively grouped into a first historical sequence, and the noise level of each data in each first historical sequence is obtained according to the data change difference in each first historical sequence; According to the time length of the current period, the first historical sequence corresponding to each other user is divided into at least one first historical subsequence, and the power load data of the target user at each moment in the current period is combined into a current load sequence; For any first historical subsequence, obtaining a similarity index between the first historical subsequence and the current load sequence based on the noise level of each data in the first historical subsequence and the data similarity between the first historical subsequence and the current load sequence; The similarity indices corresponding to all first historical subsequences are obtained, and the first historical subsequences corresponding to all similarity indices greater than a preset similarity index threshold are recorded as reference historical data sequences.

4. The method for predicting electricity consumption behavior of electric energy meter users according to claim 3, characterized in that: Obtaining the noise level of each data in each first historical sequence according to the data change difference in each first historical sequence includes: For any data in any first historical sequence, obtain neighboring data at two moments closest to the moment corresponding to the any data; The absolute values ​​of the differences between any data and the neighborhood data at the two moments are calculated respectively to obtain a first degree of difference and a second degree of difference, and the sum of the first degree of difference and the second degree of difference is linearly normalized to obtain the noise degree of any data.

5. The method for predicting electricity consumption behavior of electric energy meter users according to claim 3, characterized in that: Obtaining a similarity index between any first historical subsequence and the current load sequence based on the noise level of each data in any first historical subsequence and the data similarity between any first historical subsequence and the current load sequence includes: For any data in any first historical subsequence, according to the position of the any data in the any first historical subsequence, obtaining target data having the same position as the any data in the current load sequence, calculating an absolute value of a difference between the any data and the target data, and calculating the inverse of a sum of the absolute value of the difference and a preset constant to obtain a degree of similarity between the any data and the target data; Accumulating the noise levels of all data in any first historical subsequence to obtain an overall noise level of any first historical subsequence, calculating a ratio of the noise level of any data to the overall noise level, and subtracting the ratio from a constant 1 to obtain a similarity credibility between any data and the target data; Using the similarity credibility between the any data and the target data as the weight of the similarity degree, weighting the similarity degree to obtain the weighted similarity between the any data and the target data; The weighted similarities corresponding to all data in any first historical subsequence are calculated, and the accumulated values ​​of all weighted similarities are linearly normalized to obtain a similarity index between any first historical subsequence and the current load sequence.

6. The method for predicting electricity consumption behavior of electric energy meter users according to claim 1, characterized in that: Obtaining the degree of correlation between any user feature and the user's electricity usage behavior based on the similarity of changes in the power load data of each user in each category before the current period includes: For any category, denoising the power load data of each user in the category within a preset time range before a current period is performed to obtain a second historical sequence of each user in the category, and dividing each second historical sequence into at least one second historical subsequence based on the length of the current period; Calculating the mean of all data in each of the second historical subsequences respectively, performing linear normalization on the mean values ​​corresponding to all the second historical subsequences, and obtaining the load mean value for the period of each of the second historical subsequences; Performing linear fitting on the data in each of the second historical subsequences to obtain a fitted straight line for each of the second historical subsequences, calculating the slope of the fitted straight line for each of the second historical subsequences, and linearly normalizing all the slopes to obtain a load trend value for the time period of each of the second historical subsequences; Constructing a scatter plot based on the load mean and load trend value of all second historical subsequences under any category, where the abscissa of the scatter plot represents the load mean and the ordinate represents the load trend value; calculating the Euclidean distance between every two data points in the scatter plot, summing the reciprocals of all Euclidean distances to obtain a cumulative value, and linearly normalizing the cumulative value to obtain a power usage similarity index for all users under any category; The average value of the electricity usage similarity index of all categories under the any user feature is calculated to obtain the degree of association between the any user feature and the user's electricity usage behavior.

7. The method for predicting electricity consumption behavior of electric energy meter users according to claim 1, characterized in that: The method of obtaining target categories of the target user and other users corresponding to each reference historical data sequence, and obtaining predicted load data of the target user at each moment in the future period based on the degree of association corresponding to the user characteristics of each target category and the power load data in a period after the period of each reference historical data sequence, includes: For any reference historical data sequence, if any category contains both other users corresponding to the reference historical data sequence and the target user, then the category is recorded as the target category, all target categories are obtained, and the user features belonging to all target categories are recorded as target user features; Set the similarity value between the target user and other users corresponding to any reference historical data sequence under each target user feature to be a constant of 1, and set the similarity value between the target user and other users corresponding to any reference historical data sequence under each non-target user feature to be a constant of 0; The similarity value corresponding to each user feature is used as the weight of the corresponding correlation degree of each user feature, and the weighted average of the corresponding correlation degrees of all user features is calculated to obtain the electricity consumption similarity between other users corresponding to any reference historical data sequence and the target user; Obtaining the degree to which any reference historical data sequence reflects the target user's electricity usage behavior in the current time period based on environmental differences between the time period corresponding to any reference historical data sequence and the current time period, and the similarity between electricity usage of other users corresponding to any reference historical data sequence and the target user; The power load data in a period after the period in which each reference historical data sequence is located are respectively combined into a target historical sequence. According to the data in each target historical sequence and the reflection degree of the reference historical data sequence corresponding to each target historical sequence, the predicted load data of the target user at each moment in the future period is obtained.

8. The method for predicting electricity consumption behavior of electric energy meter users according to claim 7, characterized in that: Obtaining the degree to which any reference historical data sequence reflects the target user's electricity usage behavior in the current time period based on the environmental differences between the time period corresponding to any reference historical data sequence and the current time period, and the electricity usage similarity between other users corresponding to any reference historical data sequence and the target user, includes: Obtain the historical temperature average and the historical humidity average within the corresponding time period of any reference historical data sequence, as well as the real-time temperature average and the real-time humidity average within the current time period; Calculating the absolute value of the difference between the historical temperature mean and the real-time temperature mean to obtain a temperature difference, calculating the reciprocal of a preset constant and the temperature difference to obtain temperature similarity, obtaining humidity similarity based on the historical humidity mean and the real-time humidity mean, and linearly normalizing the sum of the temperature similarity and the humidity similarity to obtain environmental similarity between a corresponding time period of any reference historical data sequence and the current time period; The product of the electricity usage similarity between other users and the target user corresponding to any reference historical data sequence and the environmental similarity is calculated to obtain the degree to which any reference historical data sequence reflects the electricity usage behavior of the target user in the current time period.

9. The method for predicting electricity consumption behavior of electric energy meter users according to claim 7, characterized in that: Obtaining the predicted load data of the target user at each moment in the future period based on the data in each target historical sequence and the reflection degree of the reference historical data sequence corresponding to each target historical sequence includes: Calculate the cumulative value of the reflection degree of the reference historical data series corresponding to all target historical series to obtain a comprehensive reflection degree, and calculate the proportion of the reflection degree of the reference historical data series corresponding to each target historical series in the comprehensive reflection degree to obtain a reflection weight of each target historical series; Each moment of the target user in the future time period is organized into a time series. For any moment in the time series, according to the position of any moment in the time series, target historical data with the same position as that of any moment is obtained in each target historical sequence. Each target historical data is weighted according to the reflection weight of the target historical sequence in which each target historical data is located to obtain a weighted average value of all target historical data. The weighted average value is used as the predicted load data of the target user at any moment in the future time period.

10. The method for predicting electricity consumption behavior of electric energy meter users according to claim 1, characterized in that: The user characteristics include the user's electricity usage type, the user's house area, the user's floor, and the user's area.

Images