Power sale prediction method based on big data analysis

Abstract

The application relates to the technical field of data processing, in particular to a power sale prediction method based on big data analysis, which comprises the following steps: obtaining historical power sale data; obtaining a power sale quantity fitting curve based on the historical power sale data, obtaining the power sale quantity fluctuation degree of each month; obtaining an initial sequence and a plurality of difference sequences of each month; obtaining the stationary degree of each difference sequence of each month, and then obtaining the power sale quantity stationary degree of each month in each year; obtaining the stationary degree sequence of each month in each year; obtaining the importance of each month to a prediction month, and then obtaining an updated importance degree sequence of the prediction month; obtaining the power sale quantity adjustment weight of each month, adjusting an ARIMA model, combining the power sale quantity of each month, and obtaining a power sale quantity prediction value of the prediction month. The application can analyze historical power sale data, adjust ARIMA model parameters, and obtain a predicted power sale quantity with higher reference value.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically to a method for forecasting electricity sales based on big data analysis. Background Technology

[0002] The power industry is a capital-intensive sector. By forecasting electricity sales, it's possible to better plan power generation, transmission, and distribution resources, improve resource utilization, and formulate more targeted marketing and pricing strategies. Therefore, monthly electricity sales forecasting is necessary. ARIMA models are commonly used for this purpose; however, ARIMA models, which use raw data weights, have limitations when handling non-stationary data. They ignore the cyclical changes and long-term growth trends in electricity sales data, resulting in forecasts that fail to incorporate historical fluctuations and thus affecting the accuracy of the predicted electricity sales figures. Summary of the Invention

[0003] This invention provides a method for electricity sales forecasting based on big data analysis to solve existing problems.

[0004] The electricity sales forecasting method based on big data analysis of the present invention adopts the following technical solution:

[0005] This invention proposes a method for electricity sales forecasting based on big data analysis, which includes the following steps:

[0006] Obtain historical electricity sales data; the historical electricity sales data includes the electricity sales volume for several months over several years;

[0007] Based on historical electricity sales data, a fitting curve for electricity sales is obtained, and then several extreme points on the fitting curve are obtained, and several segmented curves for electricity sales are obtained. Based on the segmented curves for electricity sales and the electricity sales of the same month in different years, the degree of fluctuation in electricity sales for each month is obtained.

[0008] Based on the fluctuation of electricity sales in adjacent months, the initial sequence and several corresponding difference sequences for each month are obtained; based on the numerical differences and changes of data points in the difference sequences, the stability of each difference sequence for each month is obtained, and thus the stability of electricity sales for each month in the year is obtained.

[0009] Obtain the adjacent reference months for each month, and combine them with the stability of electricity sales for each month in the year to obtain the stability sequence for each month in the year; based on the stability sequence of each month and the stability sequence of the predicted month, obtain the importance of each month to the predicted month, and then obtain the updated importance sequence of the predicted month.

[0010] Based on the update importance sequence, the monthly electricity sales adjustment weight is obtained, and the ARIMA model is adjusted accordingly. Combined with the monthly electricity sales, the predicted electricity sales value for the forecast month is obtained.

[0011] Furthermore, the specific steps involved in obtaining a power sales fitting curve based on historical power sales data, then obtaining several extreme points on the power sales fitting curve, and dividing it into several power sales segment curves are as follows:

[0012] A coordinate system is constructed with the month as the horizontal axis and the monthly electricity sales volume as the vertical axis. The electricity sales volume of each month in the historical electricity sales data is converted into coordinate points in the coordinate system. By connecting the coordinate points of adjacent months, the electricity sales volume fitting curve of the power company is obtained.

[0013] All extreme points of the fitted curve of electricity sales volume are obtained, and the fitted curve of electricity sales volume is segmented according to the extreme points to obtain several segmented curves of electricity sales volume.

[0014] Furthermore, the specific calculation formula for obtaining the monthly fluctuation of electricity sales based on the segmented electricity sales curve and the electricity sales of the same month in different years is as follows:

[0015]

[0016] Among them, B i Z represents the fluctuation level of electricity sales in the i-th month; N represents the number of years in which electricity sales occurred; n,i Z represents the electricity sales volume in the i-th month of the nth year; n,i,min Z represents the minimum value on the piecewise curve of electricity sales in the i-th month of the nth year; n,i,max Z represents the minimum value on the piecewise curve of electricity sales in the i-th month of the nth year; n+1,i Let || represent the electricity sales volume in the i-th month of the (n+1)-th year, and || represent the absolute value function.

[0017] Furthermore, the specific steps for obtaining the initial sequence and corresponding differential sequences for each month based on the fluctuations in electricity sales between adjacent months are as follows:

[0018] On the electricity sales curve, taking one month in any year as the center, several months within the first radius of that month are recorded as reference months. The fluctuation of electricity sales between that month and all the reference months is used to form the initial sequence of that month in the order of the months.

[0019] Iteratively differencing the initial sequence for that month yields all differencing sequences for that month.

[0020] Furthermore, the stationarity of each difference sequence for each month is obtained based on the numerical differences and changes of data points in the difference sequence, and the specific calculation formula is as follows:

[0021]

[0022] Among them, P j,x R represents the stationarity of the x-th difference sequence in the j-th month; j,x B represents the number of data points in the x-th difference sequence of the j-th month; ′ j,x,r This represents the value of the r-th data point in the x-th difference sequence of the j-th month; exp() represents the mean of all data points in the x-th difference sequence in the j-th month; exp() is an exponential function with the natural constant as the base.

[0023] Furthermore, the specific steps involved in obtaining the stability of electricity sales volume for each month of the year are as follows:

[0024] For any given month, the stationarity of all difference sequences is linearly normalized. Among the difference sequences in that month whose normalized stationarity is greater than the stationarity threshold, the stationarity of the difference sequence with the smallest difference order is taken as the stationarity of electricity sales in that month.

[0025] Furthermore, the specific steps involved in obtaining the adjacent reference months for each month and combining them with the stability of electricity sales for each month throughout the year to obtain the stability sequence for each month throughout the year are as follows:

[0026] Starting from any month other than the first number of months in the electricity sales fitting curve, the stability of electricity sales of all adjacent reference months of that month is used to construct a stability sequence for that month in the order of the months.

[0027] Furthermore, the specific steps for obtaining the adjacent reference month for each month are as follows:

[0028] On the electricity sales fitting curve, taking any month other than the first number of months in the electricity sales fitting curve as the starting point, several months within the first number of months before that month are recorded as the adjacent reference months of that month.

[0029] Furthermore, the specific steps involved in obtaining the importance of each month to the predicted month based on the stationarity sequence of each month and the stationarity sequence of the predicted month, and thus obtaining the updated importance sequence of the predicted month, are as follows:

[0030]

[0031] Among them, G jIndicates the importance of month j to the predicted month; R represents the second quantity; P j,r P represents the stability of the r-th month among the adjacent reference months of the j-th month; r P represents the stability of the r-th month among the adjacent reference months of the predicted month; j Let represent the stationarity sequence for the j-th month; P represents the stationarity sequence for the predicted month; Cov() represents the correlation function; exp() is an exponential function with the natural constant as the base.

[0032] The importance of all months other than the first number of months to the predicted month is used to construct an updated importance sequence for the predicted month, arranged in the order of the months.

[0033] Furthermore, the monthly electricity sales adjustment weight is obtained based on the updated importance sequence, and the specific calculation formula is as follows:

[0034] W k =sigmoid(w k ×(1+G k ))

[0035] Among them, W k This indicates the weighting adjustment for electricity sales in month k; w k G represents the initial weight of electricity sales in the k-th month; k This indicates the importance of the k-th month to the predicted month; sigmoid() is the sigmoid function.

[0036] The beneficial effects of the technical solution of the present invention are:

[0037] By statistically analyzing historical electricity sales data, the accuracy and reference value of electricity sales forecasts can be effectively improved. Specifically, by obtaining a fitted curve of electricity sales based on historical data and identifying extreme points in electricity sales changes, the electricity sales curve can be effectively segmented, allowing for a more refined analysis of the degree of fluctuation in electricity sales and a more accurate reflection of the cyclical changes in electricity sales. Then, by calculating the degree of fluctuation in electricity sales each month, an initial sequence and corresponding difference sequence for each month are obtained. Using the analysis of the stationarity of the difference sequence, the stationarity of electricity sales for each month in each year is obtained, providing a more stable data foundation for electricity sales forecasting and thus improving the reliability of the forecast. Using the stationarity sequence, combined with adjacent reference months for each month, the importance of each month to the forecast month can be accurately assessed, constructing an updated importance sequence for the forecast month. This ensures that the contribution of each month's electricity sales data to the forecast is dynamically adjusted according to actual fluctuations, thereby improving the adaptability and flexibility of the model. Finally, by adjusting the parameters in the ARIMA model to better adapt to the periodic changes and long-term growth trends of electricity sales data, and by using the updated importance sequence to obtain the monthly electricity sales adjustment weights, a more accurate prediction value is obtained by combining the electricity sales data. This makes the ARIMA model more flexible in handling non-stationary data and solves the shortcomings of the traditional ARIMA model in handling periodic changes and long-term trends. Attached Figure Description

[0038] To more clearly illustrate the technical solutions 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.

[0039] Figure 1 This is a flowchart illustrating the steps of the electricity sales forecasting method based on big data analysis according to the present invention. Detailed Implementation

[0040] 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 the electricity sales forecasting method based on big data analysis 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.

[0041] 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.

[0042] The specific solution of the electricity sales forecasting method based on big data analysis provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0043] Please see Figure 1 The diagram illustrates a flowchart of a power sales forecasting method based on big data analysis according to an embodiment of the present invention. The method includes the following steps:

[0044] Step S001: Obtain historical electricity sales data.

[0045] The main purpose of this embodiment is to predict the electricity sales volume of the power company, which is to analyze the trend of data changes by using the historical electricity sales data of the power company.

[0046] Specifically, this embodiment collects the electricity sales data of the power company for each month over the past ten years and performs two preprocessing operations: data cleaning and data integration. The data is then unified as historical electricity sales data, and all the collected electricity sales data after preprocessing is recorded as historical electricity sales data.

[0047] Step S002: Obtain the electricity sales fitting curve based on historical electricity sales data, and then obtain several extreme points on the electricity sales fitting curve, and divide it into several electricity sales segment curves; according to the electricity sales segment curves and the electricity sales of the same month in different years, obtain the degree of fluctuation of electricity sales in each month.

[0048] It should be noted that the ARIMA model performs poorly when dealing with non-stationary data using the raw weights of historical electricity sales. The main reasons for data non-stationarity are seasonal variations in electricity consumption and overall increases in electricity consumption due to long-term economic growth. Therefore, it is necessary to first analyze the seasonal variations and overall trend characteristics of historical electricity sales to understand the degree of fluctuation in electricity sales across different months. Then, based on the degree of fluctuation in different months, determine the importance of historical electricity sales to the current predicted monthly electricity sales. Finally, optimize model performance by adjusting the weights of historical electricity sales in the ARIMA model.

[0049] It should be further noted that electricity sales are related to the current season and weather conditions. Electricity consumption in summer and winter is higher than in spring and autumn. Electricity sales will have corresponding seasonal or cyclical changes every year. At the same time, with economic growth, users' electricity consumption will also increase. The weight of historical electricity sales in the original model ignores these fluctuations. To adjust the weight of historical electricity sales in the model, the degree of fluctuation in electricity sales needs to be considered first.

[0050] Specifically, a coordinate system is constructed with the month as the horizontal axis and the monthly electricity sales volume as the vertical axis. The electricity sales volume for each month of each year in the historical electricity sales data is converted into coordinate points in the coordinate system. Connecting the coordinate points of adjacent months yields the electricity sales volume fitting curve for the power company. All extreme points of the fitted curve are obtained, and the curve is segmented based on these extreme points to obtain several segmented electricity sales volume curves. The specific formula for calculating the fluctuation of monthly electricity sales volume is as follows:

[0051]

[0052] Among them, B i Z represents the fluctuation level of electricity sales in the i-th month; N represents the number of years in which electricity sales occurred; n,i Z represents the electricity sales volume in the i-th month of the nth year; n,i,min Z represents the minimum value on the piecewise curve of electricity sales in the i-th month of the nth year; n,i,max Z represents the minimum value on the piecewise curve of electricity sales in the i-th month of the nth year; n+1,i Let || represent the electricity sales volume in the i-th month of the (n+1)-th year, and || represent the absolute value function.

[0053] It should be noted that |Z n,i -Z n,i,min | represents the degree of difference between the electricity sales volume in the i-th month of the nth year and the minimum value of that month on the piecewise electricity sales curve; |Z n,i,max -Z n,i,min | represents the degree of difference between the maximum and minimum values ​​on the piecewise curve of electricity sales in the i-th month of the n-th year; This represents the position of electricity sales in the i-th month of year n within the curve segment, measuring the degree of fluctuation in electricity sales during seasonal variations. A larger value indicates greater fluctuation in electricity sales. |Z n+1,i -Z n,i | indicates the degree of difference between the electricity sales in the (n+1)th year and the electricity sales in the i-th year; This represents the average growth rate of electricity sales in the i-th month over the years, reflecting the change in electricity sales in that month within the overall long-term trend. The larger the value, the greater the change in electricity sales in the overall long-term trend, and the greater the degree of fluctuation.

[0054] Step S003: Based on the fluctuation of electricity sales in adjacent months, obtain the initial sequence and several corresponding difference sequences for each month; based on the numerical differences and changes of data points in the difference sequences, obtain the stability of each difference sequence for each month, and then obtain the stability of electricity sales for each month in the year.

[0055] It should be noted that the ARIMA model requires differencing to stabilize the data. Prediction is made based on the stable data sequence. By differencing the initial sequence, as the order increases, a difference sequence is obtained that is the number of months minus one in the initial sequence. The stationarity of multiple difference sequences is determined. The stability of the electricity sales is obtained based on the degree of fluctuation of the electricity sales sequence after differencing.

[0056] Specifically, the first radius is preset to 3, and this is used as an example for description. Other values ​​can be set in other embodiments. On the electricity sales fitting curve, taking one month in any year as the center, several months within the first radius of that month are recorded as reference months. The electricity sales fluctuation of that month and all reference months are arranged in the order of the months to form the initial sequence of that month. Iterative difference is performed on the initial sequence of that month to obtain all difference sequences of that month. Iterative difference is a well-known technique to obtain several difference sequences from the sequence. The order of each difference is different, and the specific process will not be described in this embodiment. Based on the value and number of all data points in each difference sequence of each month, the stationarity of each difference sequence in each month is obtained. The specific calculation formula is as follows:

[0057]

[0058] Among them, P j,x R represents the stationarity of the x-th difference sequence in the j-th month; j,x B represents the number of data points in the x-th difference sequence of the j-th month; ′ j,x,r This represents the value of the r-th data point in the x-th difference sequence of the j-th month; This represents the mean of all data points in the x-th difference sequence in the j-th month; exp() is an exponential function with the natural constant as the base. In this embodiment, the exp(-x) model is used to present the inverse proportional relationship, where x is the input of the model. The implementer can set the inverse proportional function according to the actual situation.

[0059] It should be noted that, This represents the consistency of the volatility of the x-th difference sequence in the j-th month. The smaller the consistency, the more consistent the volatility of the elements in the difference sequence. At the same time, the smaller the mean of all data points in the x-th difference sequence in the j-th month, the smaller the volatility of the elements in the difference sequence. The volatility is the change in electricity sales in the difference sequence. The more consistent the change and the smaller the mean, the higher the stability of the elements in the difference sequence.

[0060] Furthermore, a preset stability threshold of 0.9 is used as an example for description; other values ​​can be set in other implementations. The stability of all difference sequences in any given month is linearly normalized. Among the difference sequences in that month whose normalized stability is greater than the stability threshold, the stability of the difference sequence with the smallest difference order is taken as the stability of electricity sales for that month.

[0061] Step S004: Obtain the adjacent reference month for each month, and combine it with the stability of electricity sales in each month of the year to obtain the stability sequence of each month in the year; based on the stability sequence of each month and the stability sequence of the predicted month, obtain the importance of each month to the predicted month, and then obtain the updated importance sequence of the predicted month.

[0062] It should be noted that the importance of each month's electricity sales is determined based on the stability of that month's sales, and the weights are adjusted accordingly. The smaller the difference between the stability of electricity sales in the preceding months of a historical month and the stability of electricity sales in the preceding months of the forecast month, the higher the reference value of that month's electricity sales for the forecast month's electricity sales.

[0063] Specifically, the first quantity is preset to 5, and this is used as an example for description. Other values ​​can be set in other implementations. On the electricity sales fitting curve, taking any month other than the months with the first quantity in the electricity sales fitting curve as the starting point, several months within the first quantity before that month are recorded as the adjacent reference months. The stability of electricity sales of all adjacent reference months is used to form the stability sequence of that month in the order of the months. It should be noted that in subsequent calculations, the months with the first quantity of historical electricity sales data are not included in the subsequent calculations. The next month after the current month is used as the prediction month. Similarly, based on the stability of electricity sales of the most recent first quantity month, the stability sequence of the prediction month is obtained. The specific calculation formula for the importance of each month to the prediction month is as follows:

[0064]

[0065] Among them, G j Indicates the importance of month j to the predicted month; R represents the second quantity; P j,r P represents the stability of the r-th month among the adjacent reference months of the j-th month; r P represents the stability of the r-th month among the adjacent reference months of the predicted month; j denoted as the stationarity sequence of the j-th month; P represents the stationarity sequence of the predicted month; Cov() represents the correlation function; exp() is an exponential function with the natural constant as the base. In this embodiment, the exp(-x) model is used to present the inverse proportional relationship and normalization processing, where x is the input of the model. Implementers can set the inverse proportional function and normalization function according to the actual situation.

[0066] It should be noted that, This represents the squared difference between the stationarity of the adjacent reference months of month j and the adjacent reference months of the predicted month. It reflects the degree of difference between the two series. The smaller the degree of difference, the smaller the difference between the stationarity, the more meaningful it is for the electricity sales of the predicted month, and the higher its importance for the electricity sales of the predicted month. This indicates a normalization operation on the degree of difference to maintain the same order of magnitude; Cov(P) j P) represents the correlation between the stationarity sequence of monthly electricity sales in the neighborhood of the j-th month and the stationarity sequence of monthly electricity sales in the neighborhood of the predicted month. It reflects the influence of the historical stationarity sequence of electricity sales on the stationarity sequence of the predicted month. The larger the value, the greater the importance to the predicted month's electricity sales.

[0067] Furthermore, the importance of all months other than the first number of months to the predicted month is used to construct an updated importance sequence for the predicted month in order of month.

[0068] Step S005: Based on the update importance sequence, obtain the monthly electricity sales adjustment weight, adjust the ARIMA model, and combine the monthly electricity sales to obtain the predicted electricity sales value for the predicted month.

[0069] It should be noted that, based on the updated importance sequence, the initial weights of the electricity sales volume for the corresponding historical months in the ARIMA model are adjusted to obtain the adjusted weights for each month's electricity sales volume. The initial weights for electricity sales volume are existing parameters in the ARIMA model, which will not be elaborated further in this embodiment. The specific formula is as follows:

[0070] W k =sigmoid(w k ×(1+G k ))

[0071] Among them, W k This indicates the weighting adjustment for electricity sales in month k; w k G represents the initial weight of electricity sales in the k-th month; k This indicates the importance of the k-th month to the predicted month; sigmoid() is the sigmoid function, which is used in this embodiment for normalization.

[0072] It should be noted that w k ×(1+G k The expression indicates that the weight of the electricity sales in the k-th month in the ARIMA model is adjusted based on the importance of the electricity sales in the k-th month to the electricity sales in the predicted month, thus obtaining the electricity sales adjustment weight.

[0073] Furthermore, by adjusting the weights based on the monthly electricity sales, an adjusted ARIMA model is obtained. By inputting the monthly electricity sales data from the historical sales data for each year, the electricity sales for the predicted month are predicted, and the predicted electricity sales value for the predicted month is obtained, thus completing the prediction of the electricity sales for the power company in a given month.

[0074] This concludes the embodiment.

[0075] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for electricity sales forecasting based on big data analysis, characterized in that, The method includes the following steps: Obtain historical electricity sales data; the historical electricity sales data includes the electricity sales volume for several months over several years; Based on historical electricity sales data, a fitting curve for electricity sales is obtained, and then several extreme points on the fitting curve are obtained, and several segmented curves for electricity sales are obtained. Based on the segmented curves for electricity sales and the electricity sales of the same month in different years, the degree of fluctuation in electricity sales for each month is obtained. Based on the fluctuation of electricity sales in adjacent months, the initial sequence and several corresponding difference sequences for each month are obtained; based on the numerical differences and changes of data points in the difference sequences, the stability of each difference sequence for each month is obtained, and thus the stability of electricity sales for each month in the year is obtained. Obtain the adjacent reference months for each month, and combine them with the stability of electricity sales for each month in the year to obtain the stability sequence for each month in the year; based on the stability sequence of each month and the stability sequence of the predicted month, obtain the importance of each month to the predicted month, and then obtain the updated importance sequence of the predicted month. Based on the update importance sequence, the monthly electricity sales adjustment weight is obtained, and the ARIMA model is adjusted accordingly. Combined with the monthly electricity sales, the predicted electricity sales value for the forecast month is obtained.

2. The electricity sales forecasting method based on big data analysis according to claim 1, characterized in that, The specific steps involved in obtaining a power sales fitting curve based on historical power sales data, then identifying several extreme points on the fitting curve, and finally dividing the curve into several segmented power sales curves are as follows: A coordinate system is constructed with the month as the horizontal axis and the monthly electricity sales volume as the vertical axis. The electricity sales volume of each month in the historical electricity sales data is converted into coordinate points in the coordinate system. By connecting the coordinate points of adjacent months, the electricity sales volume fitting curve of the power company is obtained. All extreme points of the fitted curve of electricity sales volume are obtained, and the fitted curve of electricity sales volume is segmented according to the extreme points to obtain several segmented curves of electricity sales volume.

3. The electricity sales forecasting method based on big data analysis according to claim 1, characterized in that, The fluctuation level of monthly sales volume is obtained based on the segmented sales volume curve and the sales volume of the same month in different years. The specific calculation formula is as follows: in, Indicates the first The degree of fluctuation in monthly electricity sales; Indicates the number of years in which electricity was sold; Indicates the first Year Monthly electricity sales volume; Indicates the first Year The minimum value of monthly electricity sales on the segmented curve of electricity sales; Indicates the first Year The maximum value of monthly electricity sales on the segmented curve of electricity sales; Indicates the first Year Electricity sales volume for the month This represents the absolute value function.

4. The electricity sales forecasting method based on big data analysis according to claim 1, characterized in that, The specific steps for obtaining the initial sequence and corresponding difference sequences for each month based on the fluctuation of electricity sales in adjacent months are as follows: On the electricity sales curve, taking one month in any year as the center, several months within the first radius of that month are recorded as reference months. The fluctuation of electricity sales between that month and all the reference months is used to form the initial sequence of that month in the order of the months. Iteratively differencing the initial sequence for that month yields all differencing sequences for that month.

5. The electricity sales forecasting method based on big data analysis according to claim 1, characterized in that, The stationarity of each difference sequence for each month is obtained based on the numerical differences and changes of data points in the difference sequence. The specific calculation formula is as follows: in, Indicates the first The first month The stationarity of the difference sequence; Indicates the first The first month The number of data points in each difference sequence; Indicates the first The first month In the nth difference sequence The value of each data point; Indicates the first The first month The mean of all data points in a difference sequence; It is an exponential function with the natural constant as its base.

6. The electricity sales forecasting method based on big data analysis according to claim 1, characterized in that, The specific steps involved in obtaining the stability of electricity sales volume for each month of the year are as follows: For any given month, the stationarity of all difference sequences is linearly normalized. Among the difference sequences in that month whose normalized stationarity is greater than the stationarity threshold, the stationarity of the difference sequence with the smallest difference order is taken as the stationarity of electricity sales in that month.

7. The electricity sales forecasting method based on big data analysis according to claim 1, characterized in that, The specific steps involved in obtaining the adjacent reference months for each month and combining them with the stability of electricity sales for each month throughout the year to obtain the stability sequence for each month throughout the year are as follows: Starting from any month other than the first number of months in the electricity sales fitting curve, the stability of electricity sales of all adjacent reference months of that month is used to construct a stability sequence for that month in the order of the months.

8. The electricity sales forecasting method based on big data analysis according to claim 7, characterized in that, The specific steps for obtaining the adjacent reference month for each month are as follows: On the electricity sales fitting curve, taking any month other than the first number of months in the electricity sales fitting curve as the starting point, several months within the first number of months before that month are recorded as the adjacent reference months of that month.

9. The electricity sales forecasting method based on big data analysis according to claim 1, characterized in that, The specific steps involved in obtaining the importance of each month to the predicted month based on the stationarity sequence of each month and the stationarity sequence of the predicted month, and then obtaining the updated importance sequence of the predicted month, are as follows: in, Indicates the first The importance of the month in predicting the month; Indicates the second quantity; Indicates the first The adjacent reference month of the month The degree of stability over the past month; This indicates the first reference month adjacent to the predicted month. The degree of stability over the past month; Indicates the first A series of months with stable performance; This represents a sequence indicating the stationarity of the predicted months. Represents the relevant function; It is an exponential function with the natural constant as its base; The importance of all months other than the first number of months to the predicted month is used to construct an updated importance sequence for the predicted month, arranged in the order of the months.

10. The electricity sales forecasting method based on big data analysis according to claim 1, characterized in that, The monthly electricity sales adjustment weight is obtained based on the update importance sequence, and the specific calculation formula is as follows: in, Indicates the first Weighting of monthly electricity sales volume; Indicates the first The initial weight of monthly electricity sales volume; Indicates the first The importance of each month to the month of prediction; for function.

Images