Electric energy prediction method and system based on multi-factor correlation coefficient, storage medium and equipment

Abstract

The invention discloses an electric energy prediction method and system based on a multi-factor correlation coefficient, a storage medium and equipment. The method comprises the following steps: acquiring historical electricity consumption, temperature and date type of a certain region; dividing upper and lower limits of temperature intervals according to temperature characteristics of each season, and calculating a correlation coefficient of historical electricity consumption and temperature of each temperature interval based on a percentile method; taking the historical electricity consumption of the workday as a reference value, performing normalization processing on the historical electricity consumption of each date type based on the reference value, and calculating a correlation coefficient between the historical electricity consumption of each date type and the date type; according to the method, the correlation coefficient of the historical electricity consumption and the temperature and the correlation coefficient of the historical electricity consumption and the date type are used as features and are input into a regional electricity quantity prediction model XG-Boost, the electric energy in a certain time period in the future is predicted, the electric energy prediction precision is effectively improved, and accurate guidance is provided for power supply of a power grid.

Description

Technical Field

[0001] The present invention relates to the technical field of electric energy prediction, and specifically, to an electric energy prediction method, system, storage medium, and device based on multi-factor correlation coefficients. Background Art

[0002] Electric energy prediction refers to the work of estimating the electric energy consumption situation in a future time period by analyzing historical electric energy data and related influencing factors and applying appropriate mathematical models and algorithms. Electric energy prediction technologies can be classified into statistical analysis methods, regression analysis methods, machine learning methods, deep learning methods, etc. according to the prediction methods. When performing prediction, electric energy prediction technologies use a variety of features, which can reflect the variation law of electric energy and its association with other factors from different aspects, mainly including week, month, temperature, humidity, wind speed, macro economy, electricity consumption scale, industrial structure, etc.

[0003] Currently, most models use machine learning or statistical analysis methods to directly use temperature and other related factors, such as time, day of the week, holidays, etc. as input features to construct an electric energy prediction model, without fully exploring the correlation relationship between temperature, holidays and electricity consumption. However, both temperature and holidays are key factors affecting power demand: a slight change in temperature may have a significant impact on the electricity consumption of residents and enterprises. If the correlation relationship between temperature and electricity consumption is not fully explored, the electric energy prediction result is likely to deviate from the actual value, resulting in a decrease in prediction accuracy; during holidays, the electricity consumption of residents may increase, and at the same time, the electricity consumption of some large industrial users may decrease due to the suspension and resumption of work during holidays. If these holiday effects are not considered, the prediction result of electric energy will also deviate from the actual value.

[0004] Therefore, there is an urgent need to construct an electric energy prediction method that can comprehensively consider the multi-factor correlation such as temperature, holidays, etc., improve the accuracy of electric energy prediction, and provide more accurate guidance for power grid power supply. Summary of the Invention

[0005] Aiming at the problems existing in the prior art, the present invention provides an electric energy prediction method, system, storage medium, and device based on multi-factor correlation coefficients. By exploring the correlation relationship between key features such as temperature, date type, etc. and electricity consumption, and using the multi-factor correlation coefficient as an input feature to perform electric energy prediction, the accuracy of electric energy prediction is effectively improved, and accurate guidance is provided for power grid power supply.

[0006] To achieve the above technical purpose, the present invention adopts the following technical solutions: An electric energy prediction method based on multi-factor correlation coefficients specifically includes the following steps:

[0007] Step S1: Obtain the historical electricity consumption, temperature, and date type of a certain region;

[0008] Step S2: Divide the upper and lower limits of the temperature range according to the temperature characteristics of each season, and calculate the correlation coefficient between the historical electricity consumption and temperature in each temperature range based on the percentile method;

[0009] Step S3: Use the historical electricity consumption on weekdays as the benchmark value, and normalize the historical electricity consumption of each date type based on the benchmark value. Calculate the correlation coefficient between the historical electricity consumption and date type under each date type through statistical analysis methods;

[0010] Step S4: Take the correlation coefficient between the historical electricity consumption and temperature and the correlation coefficient between the historical electricity consumption and date type as features, and input them into the regional electricity consumption prediction model XG - Boost to predict the electricity energy within a certain future time period.

[0011] Further, step S2 includes the following sub - steps:

[0012] Step S2.1: Determine the benchmark value of the historical electricity consumption in each season, and normalize the historical electricity consumption in each season based on the benchmark value to obtain coefficient values;

[0013] Step S2.2: Divide the upper and lower limits of the temperature range according to the temperature characteristics of each season, calculate the mean absolute deviation rate corresponding to different percentile coefficient values within each temperature range, and determine the correlation coefficient between the historical electricity consumption and temperature in the corresponding temperature range.

[0014] Further, the process of normalizing the historical electricity consumption in each season based on the benchmark value to obtain coefficient values in step S2.1 is as follows:

[0015]

[0016] Among them, x represents the set of normalized coefficient values, X represents the set of historical daily electricity consumption in each season, X basic represents the benchmark value of the historical electricity consumption in each season, n represents the length of the historical electricity consumption set within the stable time period corresponding to each season, i represents the index of n, and X i represents the i - th historical daily electricity consumption within the stable time period corresponding to each season.

[0017] Further, the calculation process of the correlation coefficient between the historical electricity consumption and temperature in the corresponding temperature range in step S2.2 is as follows:

[0018]

[0019] Among them, Corr [t1,t2]Denote the correlation coefficient between historical electricity consumption and temperature within the temperature range [t1, t2], p represents the percentile, pmin represents the minimum percentile, pmax represents the maximum percentile, k represents the number of days within the temperature range [t1, t2], j represents the index of k, and x j Denote the j-th coefficient value within the temperature range [t1, t2], and x p Denote the coefficient value at the percentile p of the set composed of the coefficient values x within the temperature range [t1, t2].

[0020] Furthermore, the process of normalizing the historical electricity consumption of each date type based on the benchmark value in step S3 is as follows:

[0021]

[0022] Among them, e represents the result of normalizing the historical electricity consumption of a certain date type based on the benchmark value, E represents the historical electricity consumption under a certain date type, and E basic Denote the historical electricity consumption on weekdays, h represents the number of weekdays as the benchmark value, u represents the index of h, and E u Denote the historical electricity consumption on the u-th weekday.

[0023] Furthermore, the calculation process of the correlation coefficient between the historical electricity consumption and the date type under each date type in step S3 is as follows:

[0024]

[0025] Among them, Corr (datetype) Denote the correlation coefficient between the historical electricity consumption and the date type under a certain date type, m represents the length of the set of the historical electricity consumption under a certain date type, v represents the index of m, and e v Denote the result of normalizing the v-th historical electricity consumption.

[0026] Furthermore, the input features in step S4 also include: the cross features constructed by the correlation coefficient between historical electricity consumption and temperature and the correlation coefficient between historical electricity consumption and date type: select the high-temperature correlation coefficient Corr (high_temp) and the low-temperature correlation coefficient Corr (low_temp) from the correlation coefficient between historical electricity consumption and temperature according to the historical daily maximum temperature and daily minimum temperature, and combine with the correlation coefficient Corr (datetype) between historical electricity consumption and date type to construct cross features:

[0027] Coef (temp_date)1 =Corr (high_temp) *Corr (datetype)

[0028] Coef(temp_date)2 = Corr (low_temp) *Corr (datetpye) 。

[0029] Furthermore, the present invention also provides an electric energy prediction system based on multi-factor correlation coefficients, including:

[0030] A data acquisition module, configured to acquire historical power consumption, temperature, and date type of a certain region;

[0031] A temperature correlation coefficient calculation module, configured to divide the upper and lower limits of the temperature range according to the temperature characteristics of each season, and calculate the correlation coefficient between the historical power consumption and temperature of each season;

[0032] A date type correlation coefficient calculation module, configured to calculate the correlation coefficient between the historical power consumption and the date type under each date type;

[0033] An electric energy prediction module, configured to use the correlation coefficient between the historical power consumption and temperature and the correlation coefficient between the historical power consumption and the date type as features, and input them into the regional power consumption prediction model XG-Boost to predict the electric energy.

[0034] Furthermore, the present invention also provides a computer-readable storage medium storing a computer program, and the computer program causes a computer to execute the electric energy prediction method based on multi-factor correlation coefficients.

[0035] Furthermore, the present invention also provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the electric energy prediction method based on multi-factor correlation coefficients is implemented.

[0036] Compared with the prior art, the present invention has the following beneficial effects: The electric energy prediction method and system based on multi-factor correlation coefficients of the present invention mine the correlation coefficient between the historical power consumption and temperature of each temperature range through the percentile method, and mine the correlation coefficient between the historical power consumption and the date type under each date type through the statistical analysis method, and jointly use them as the input features of the electric energy prediction model. First, the percentile method can accurately determine the values of the temperature range and power consumption at different percentiles, so as to extract the correlation coefficient between them, and thus more accurately obtain the influence degree of temperature on power consumption; second, through the statistical analysis method, the influence of different date types on power consumption can be more accurately obtained, so as to improve the accuracy of power demand prediction; finally, by capturing the complex relationships between multi-factors, the generalization ability of the electric energy prediction model is enhanced, the accuracy of electric energy prediction is improved, and effective guidance is provided for power grid power supply. Description of the Drawings

[0037] Figure 1Flow chart of the electric energy prediction method based on multi - factor correlation coefficient of the present invention;

[0038] Figure 2 Schematic diagram of the electric energy prediction system based on multi - factor correlation coefficient of the present invention. Detailed implementation manners

[0039] The technical solution of the present invention will be further explained below with reference to the accompanying drawings.

[0040] As Figure 1 Flow chart of the electric energy prediction method based on multi - factor correlation coefficient of the present invention. The electric energy prediction method specifically includes the following steps:

[0041] Step S1: Obtain the historical electricity consumption, temperature, and date type of a certain area. Among them, the temperature data includes: daily maximum temperature and daily minimum temperature; the date type includes: weekdays, weekends, and holidays.

[0042] Step S2: Divide the upper and lower limits of the temperature range according to the temperature characteristics of each season, calculate the correlation coefficient between the historical electricity consumption and temperature in each temperature range based on the percentile method, and obtain the correlation coefficient value corresponding to the percentile with the minimum mean absolute deviation rate, so as to more accurately obtain the influence degree of temperature on electricity consumption. Specifically, it includes the following sub - steps:

[0043] Step S2.1: Determine the reference value of the historical electricity consumption in each season, and normalize the historical electricity consumption in each season based on the reference value to obtain the coefficient value:

[0044]

[0045] Among them, x represents the set of coefficient values after normalization, X represents the set of historical daily electricity consumption in each season, X basic represents the reference value of the historical electricity consumption in each season, n represents the length of the set of historical electricity consumption in the stable time period corresponding to each season, i represents the index of n, and X i represents the i - th historical daily electricity consumption in the stable time period corresponding to each season.

[0046] In a technical solution of the present invention, select and count the average daily electricity consumption of weekdays during the period from April 15th to April 30th of each year, and then estimate the reference values of the historical electricity consumption in spring and summer. Select the electricity consumption data of weekdays during the periods from March to May and from June to August of each year as the electricity consumption sets in spring and summer respectively; select and count the average daily electricity consumption of weekdays during the period from October 15th to October 31st of each year, and then estimate the reference values of the historical electricity consumption in autumn and winter. Select the electricity consumption data of weekdays during the periods from September to November and from December to February of the next year as the electricity consumption sets in autumn and winter respectively.

[0047] Step S2.2: Divide the upper and lower limits of the temperature range according to the temperature characteristics of each season, calculate the mean absolute deviation rate corresponding to different percentile coefficient values within each temperature range, and determine the correlation coefficient between the historical electricity consumption and temperature in the corresponding temperature range:

[0048]

[0049] Among them, Corr [t1,t2] represents the correlation coefficient between the historical electricity consumption and temperature within the temperature range [t1, t2], p represents the percentile, pmin represents the minimum percentile, pmax represents the maximum percentile, k represents the number of days within the temperature range [t1, t2], j represents the index of k, x j represents the j-th coefficient value within the temperature range [t1, t2], x p represents the coefficient value at the percentile p of the set of coefficient values x within the temperature range [t1, t2], and the coefficient value x corresponding to the percentile p p satisfies the following condition: at least p% of the data is less than or equal to x p , for example, the 75th percentile means that 75% of the data in this group is less than or equal to the value corresponding to this position. In a technical solution of the present invention, the range of the percentile is set between 25 and 90, and a coefficient value is taken every 5 percentiles and the corresponding mean absolute deviation rate is calculated. The coefficient value corresponding to the minimum mean absolute deviation rate among these percentiles is the correlation coefficient between the historical electricity consumption and temperature in the corresponding temperature range. Using this method to calculate the correlation coefficients between the historical electricity consumption and temperature in different temperature ranges for a certain user group in a certain area in summer and winter respectively, where the correlation coefficient corresponding to the daily maximum temperature range in summer and the correlation coefficient corresponding to the daily minimum temperature range in winter are taken, and the results are shown in Table 1.

[0050] Table 1 Correlation coefficients of different temperature ranges for a certain user group in a certain area

[0051]

[0052]

[0053] Step S3: Use the historical electricity consumption on weekdays as the benchmark value, and normalize the historical electricity consumption of each date type based on the benchmark value. By using statistical analysis methods to calculate the correlation coefficient between the historical electricity consumption and the date type under each date type, the impact of different date types on electricity consumption can be obtained more accurately, thereby improving the accuracy of power demand forecasting.

[0054] The process of normalizing the historical electricity consumption of each date type based on the benchmark value in the present invention is as follows:

[0055]

[0056] Among them, e represents the result of normalizing the historical electricity consumption of a certain date type based on a reference value, E represents the historical electricity consumption under a certain date type, and E basic represents the historical electricity consumption on weekdays, h represents the number of weekdays used as the reference value, u represents the index of h, and E u represents the historical electricity consumption on the u-th weekday.

[0057] In a technical solution of the present invention, the correlation coefficient of weekdays is set to 1, that is, it is used as a benchmark and does not need to be calculated. For the remaining date types, the reference electricity consumption needs to be determined separately and normalized: the reference electricity consumption on Saturdays and Sundays is determined by selecting and counting the average electricity consumption on weekdays in spring and autumn each year. The reference electricity consumption on New Year's Day, the Dragon Boat Festival, the Tomb-Sweeping Festival, Labor Day, the Mid-Autumn Festival, and National Day is determined according to the average electricity consumption on weekdays in the week before the festival. The reference electricity consumption for the Spring Festival needs to first determine the number of influencing days before and after the Spring Festival holiday, and then determine it according to the average electricity consumption on weekdays in the week before the influencing days before the festival.

[0058] The calculation process of the correlation coefficient between the historical electricity consumption and the date type under each date type in the present invention is as follows:

[0059]

[0060] Among them, Corr (datetype) represents the correlation coefficient between the historical electricity consumption and the date type under a certain date type, m represents the set length of the historical electricity consumption under a certain date type, v represents the index of m, and e v represents the result of normalizing the v-th historical electricity consumption. By calculating the correlation coefficient between the historical electricity consumption and the date type under different date types for a certain user group in a certain region by the above method, the results are shown in Table 2.

[0061] Table 2 Correlation coefficients of different date types for a certain user group in a certain region

[0062] Date type Correlation coefficient Weekday 1.00 Saturday 0.94 Sunday 0.88 New Year's Day 0.84 Spring Festival 0.78 Dragon Boat Festival 0.84 Tomb-Sweeping Day 0.86 Labor Day 0.81 Mid-Autumn Festival 0.82 National Day 0.85

[0063] Step S4: Use the correlation coefficient between the historical electricity consumption and temperature and the correlation coefficient between the historical electricity consumption and the date type as features, input them into the regional electricity consumption prediction model XG-Boost, and predict the electricity energy in a future period of time. By correlating factors such as temperature and date type, the complex relationship between multiple factors can be captured, the generalization ability of the electricity energy prediction model can be enhanced, the accuracy of electricity energy prediction can be improved, and effective guidance can be provided for power grid power supply.

[0064] In one technical solution of the present invention, the input features further include: cross features constructed from the correlation coefficient between historical electricity consumption and temperature and the correlation coefficient between historical electricity consumption and date type: selecting a high-temperature correlation coefficient Corr from the correlation coefficient between historical electricity consumption and temperature according to the historical daily maximum temperature and daily minimum temperature (high_temp) and a low-temperature correlation coefficient Corr (low_temp) , and combining with the correlation coefficient Corr between historical electricity consumption and date type (datetype) to construct cross features:

[0065] Coef (temp_date)1 = Corr (high_temp) * Corr (datetype)

[0066] Coef (temp_date)2 = Corr (low_temp) * Corr (datetype) .

[0067] By considering the interaction between temperature and date type, the present invention constructs cross features to improve the accuracy of electric energy prediction; the temperature correlation coefficient, date correlation coefficient, and cross features can also be used as inputs to the regional electricity consumption prediction model XG-Boost through different feature combinations, feature selections, etc. for electric energy prediction, capturing the correlation relationships between multiple factors, making the prediction results of electric energy more in line with reality, and providing a more accurate reference for power grid power supply.

[0068] For example Figure 2 is a schematic diagram of the electric energy prediction system based on multi-factor correlation coefficients of the present invention. The electric energy prediction system includes:

[0069] A data acquisition module for acquiring historical electricity consumption, temperature, and date type of a certain region;

[0070] A temperature correlation coefficient calculation module for calculating the upper and lower limits of the temperature range according to the temperature characteristics of each season and calculating the correlation coefficient between historical electricity consumption and temperature in each season;

[0071] A date type correlation coefficient calculation module for calculating the correlation coefficient between historical electricity consumption and date type under each date type;

[0072] An electric energy prediction module for using the correlation coefficient between historical electricity consumption and temperature and the correlation coefficient between historical electricity consumption and date type as features and inputting them into the regional electricity consumption prediction model XG-Boost to predict electric energy.

[0073] In one technical solution of the present invention, there is also provided a computer-readable storage medium storing a computer program, and the computer program causes a computer to execute the above-described electric energy prediction method based on multi-factor correlation coefficients.

[0074] In one technical solution of the present invention, there is also provided an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the above-described electric energy prediction method based on multi-factor correlation coefficients is implemented.

[0075] In the embodiments disclosed in the present application, the computer storage medium may be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. The computer storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of the computer storage medium would include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0076] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in the present application can be implemented by electronic hardware or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. A professional technician can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0077] The above are only the preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions falling within the idea of the present invention belong to the protection scope of the present invention. It should be noted that for those of ordinary skill in the technical field, several improvements and refinements made without departing from the principle of the present invention should be regarded as within the protection scope of the present invention.

Claims

1. A method for predicting electric energy based on multi-factor correlation coefficients, characterized in that: The specific steps include: Step S1, obtaining historical electricity consumption, temperature and date type of a certain area; Step S2: divide the upper and lower limits of the temperature range according to the temperature characteristics of each season, and calculate the correlation coefficient between the historical power consumption and temperature in each temperature range based on the percentile method; Step S3: Taking the historical power consumption on working days as a reference value, normalizing the historical power consumption of each date type based on the reference value, and calculating the correlation coefficient between the historical power consumption and the date type under each date type by a statistical analysis method; Step S4: The correlation coefficient between historical electricity consumption and temperature and the correlation coefficient between historical electricity consumption and date type are used as features and input into the regional electricity consumption prediction model XG-Boost to predict the electricity consumption in a certain time period in the future.

2. The electric energy prediction method based on multi-factor correlation coefficient according to claim 1 is characterized in that: Step S2 includes the following sub-steps: Step S2.1, determining a benchmark value of historical electricity consumption in each season, and normalizing the historical electricity consumption in each season based on the benchmark value to obtain a coefficient value; Step S2.2: divide the upper and lower limits of the temperature range according to the temperature characteristics of each season, calculate the average absolute deviation rate corresponding to different percentile coefficient values ​​in each temperature range, and determine the correlation coefficient between the historical power consumption and temperature in the corresponding temperature range.

3. The electric energy prediction method based on multi-factor correlation coefficient according to claim 2 is characterized in that: The process of normalizing the historical electricity consumption of each season based on the benchmark value to obtain the coefficient value in step S2.1 is: Among them, x represents the normalized coefficient value set, X represents the historical daily electricity consumption set of each season, and X basic represents the baseline value of historical electricity consumption in each season, n represents the length of the historical electricity consumption set in the stable time period corresponding to each season, i represents the index of n, and X i Represents the electricity consumption of the ith historical day in the stable time period corresponding to each season.

4. The electric energy prediction method based on multi-factor correlation coefficient according to claim 3 is characterized in that: The calculation process of the correlation coefficient between the historical power consumption and temperature in the corresponding temperature range in step S2.2 is: Among them, Corr [t1,t2] represents the correlation coefficient between historical power consumption and temperature in the temperature interval [t1, t2], p represents percentile, pmin represents minimum percentile, pmax represents maximum percentile, k represents the number of days in the temperature interval [t1, t2], j represents the index of k, and x represents the value of the index of k. j represents the jth coefficient value in the temperature interval [t1, t2], x p Represents the coefficient value at percentile p of the set of coefficient values ​​x in the temperature interval [t1, t2].

5. The electric energy prediction method based on multi-factor correlation coefficient according to claim 4 is characterized in that: The process of normalizing the historical power consumption of each date type based on the reference value in step S3 is as follows: Among them, e represents the result of normalizing the historical electricity consumption of a certain date type based on the benchmark value, E represents the historical electricity consumption under a certain date type, and E basic represents the historical electricity consumption on weekdays, h represents the number of working days as the base value, u represents the index of h, E u Represents the historical electricity consumption on the u-th working day.

6. The electric energy prediction method based on multi-factor correlation coefficient according to claim 5 is characterized in that: The calculation process of the correlation coefficient between the historical power consumption and the date type under each date type in step S3 is: Among them, Corr (datetype) represents the correlation coefficient between the historical electricity consumption under a certain date type and the date type, m represents the set length of the historical electricity consumption under a certain date type, v represents the index of m, e represents the v It represents the result of normalizing the vth historical electricity consumption.

7. The electric energy prediction method based on multi-factor correlation coefficient according to claim 6 is characterized in that: The input features in step S4 also include: the correlation coefficient between historical power consumption and temperature and the cross-feature constructed by the correlation coefficient between historical power consumption and date type: the high temperature correlation coefficient Corr is selected from the correlation coefficient between historical power consumption and temperature according to the historical daily maximum temperature and daily minimum temperature. (high_temp) and low temperature correlation coefficient Corr (low_temp) , combined with the correlation coefficient Corr between historical electricity consumption and date type (datetype) Construct a cross feature: Coef (temp_date)1 =Corr (high_temp) *Corr (datetype) Coef (temp_date)2 =Corr (low_temp) *Corr (datetype) 。 8. An electric energy prediction system based on multi-factor correlation coefficients, characterized in that: include: The data acquisition module is used to obtain the historical electricity consumption, temperature and date type of a certain area; The temperature correlation coefficient calculation module is used to divide the upper and lower limits of the temperature range according to the temperature characteristics of each season, and calculate the correlation coefficient between the historical power consumption and temperature in each season; A date type correlation coefficient calculation module is used to calculate the correlation coefficient between the historical power consumption and the date type under each date type; The electric energy prediction module is used to input the correlation coefficient between historical electricity consumption and temperature and the correlation coefficient between historical electricity consumption and date type as features into the regional electricity consumption prediction model XG-Boost to predict the electric energy.

9. A computer-readable storage medium storing a computer program, characterized in that: The computer program enables the computer to execute the electric energy prediction method based on multi-factor correlation coefficients as described in any one of claims 1-7.

10. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for predicting electric energy based on multi-factor correlation coefficients as described in any one of claims 1 to 7 is implemented.

Images