Financial index and neural network electricity consumption prediction method and system, medium and equipment

Abstract

The invention relates to the technical field of electric power, and discloses a financial index and neural network electricity consumption prediction method and system, a medium and equipment, and the method comprises the steps: obtaining the historical daily electricity consumption data of a user, extracting a CCI index, a DMI index and a BOLL index from the historical daily electricity consumption data according to the characteristics of a financial index, and carrying out the calculation of the CCI index, the DMI index and the BOLL index; combining the historical daily electricity consumption data, the CCI index, the DMI index and the BOLL index and dividing into a training set and a test set; and constructing a neural network model, training the neural network model by using the training set to obtain a trained neural network model, and inputting the test set into the trained neural network model to obtain a user power consumption prediction result. Effective features in the historical electricity consumption data of the user can be effectively extracted, and the accuracy of predicting the electricity consumption of the user is improved.

Description

Technical Field

[0001] The present invention relates to the field of electric power technology, and in particular to a method, system, medium and device for predicting electricity consumption based on financial indicators and neural networks. Background Art

[0002] User electricity consumption forecasting is a technology that predicts future electricity demand by analyzing historical electricity consumption data, environmental factors and user behavior. It can provide reliability and efficiency support for the power system. Therefore, user electricity consumption forecasting is a key technology in power grid construction and management.

[0003] When predicting the power consumption of users, time series analysis is the basis. Revealing the power consumption pattern through trend and periodic detection can provide an important reference for future power demand. With the emergence of intelligent algorithms, the application of machine learning and other methods in power consumption forecasting can capture complex nonlinear relationships, making the forecasting model more intelligent and flexible, and further improving the accuracy of forecasts. Meteorological conditions and seasonal changes are important factors affecting power consumption, so they are included in the consideration of the forecasting model. Analysis of meteorological data such as temperature and humidity can help predict peaks and troughs of power consumption more accurately, while the focus on seasonal factors makes the model more comprehensive. In addition, the advancement of smart meters and Internet of Things technologies has pushed the frontier of user power consumption forecasting. Smart meters monitor power consumption in real time and provide timely data feedback, while the connectivity of Internet of Things devices provides a richer data source for the model, enhancing the adaptability and accuracy of the model. The comprehensive application of these technologies will provide strong support for the efficient operation of the power system and the optimization of user electricity costs, and promote the energy sector towards a more sustainable and intelligent future.

[0004] However, the prior art does not have a unified approach to effectively analyze various influencing factors based on comprehensive consideration of the influencing factors of electricity consumption, and is unable to extract effective features from the user's historical electricity consumption data, resulting in low accuracy in predicting the user's electricity consumption. Summary of the invention

[0005] To this end, the technical problem to be solved by the present invention is to overcome the deficiencies in the prior art and provide a method, system, medium and device for predicting electricity consumption based on financial indicators and neural networks, which can effectively extract effective features from the user's historical electricity consumption data and improve the accuracy of user electricity consumption prediction.

[0006] In order to solve the above technical problems, the present invention provides a method for predicting electricity consumption based on financial indicators and neural networks, comprising:

[0007] Obtain the user's historical daily electricity consumption data, extract the CCI index, the DMI index, and the BOLL index from the historical daily electricity consumption data according to the characteristics of the financial indicators, combine the historical daily electricity consumption data, the CCI index, the DMI index, and the BOLL index, and divide them into a training set and a test set;

[0008] A neural network model is constructed, the training set is used to train the neural network model to obtain a trained neural network model, and the test set is input into the trained neural network model to obtain a user power consumption prediction result.

[0009] In one embodiment of the present invention, the neural network model includes a convolutional neural network and a bidirectional long short-term memory neural network.

[0010] The convolutional neural network extracts local features of the training set and the test set, uses the local features extracted by the convolutional neural network from the training set to train the bidirectional long short-term memory neural network, and inputs the local features extracted by the convolutional neural network from the prediction set into the trained bidirectional long short-term memory neural network to obtain the user power consumption prediction result.

[0011] In one embodiment of the present invention, the CCI indicator is a CCI indicator value sequence, and the CCI indicator value sequence is CCI=[CCI i ,i=1,2,...,n],CCI i is the CCI indicator value of the i-th day in CCI, CCI i The calculation method is:

[0012]

[0013] Among them, TYP is the typical value sequence of power consumption, TYP = [TYP i ,i=1,2,...,n],TYP i is the typical value of electricity consumption on the ith day, n is the total number of days, and the typical value of electricity consumption is the average of the highest single-hour electricity consumption, the lowest single-hour electricity consumption, and the average daily electricity consumption of the day; MA (TYP,n) is the n-day simple moving average sequence, It represents the simple moving average of the ith day in TYP; α is the preset coefficient, and MAD represents the mean absolute error of TYP.

[0014] In one embodiment of the present invention, the MAD is the mean absolute error of TYP with a sliding window size of n days and a sliding step size of 1 day. The calculation method of MAD is:

[0015]

[0016] In one embodiment of the present invention, the DMI index includes an electricity consumption increase index value sequence, an electricity consumption decrease index value sequence, and an average trend index sequence.

[0017] The power consumption increase index value sequence is PDI=[PDI i ,i=1,2,...,n1],PDI i is the power consumption increase index value on the i-th day, n1 is the preset number of days, PDI i The calculation method is:

[0018]

[0019] in, is the n1-day simple moving average of the rising trend value of electricity consumption on the i-th day, is the n1-day simple moving average of the actual fluctuation of electricity consumption on the i-th day;

[0020] The power consumption reduction index value sequence is MDI=[MDI i ,i=1,2,...,n1],MDI i is the electricity consumption reduction index value on the i-th day, MDI i The calculation method is:

[0021]

[0022] in, is the n1-day simple moving average of the downward trend value of electricity consumption on the i-th day;

[0023] The average trend index sequence is ADX=[ADX i ,i=1,2,...,n2],ADX i is the average trend indicator of the i-th day, n2 is the preset number of days, ADX i The calculation method is:

[0024]

[0025] in, It is the n2-day simple moving average of the trend indicator sequence on the i-th day.

[0026] In one embodiment of the present invention, the power consumption rising trend value sequence is DMP=[DMP i ,i=1,2,...,n1],DMP i is the rising trend value of electricity consumption on the i-th day, DMP i The calculation method is:

[0027]

[0028] in, is the highest single-hour electricity consumption on the ith day, is the maximum hourly electricity consumption on the i-1th day;

[0029] The real amplitude sequence of electricity consumption is TR = [TR i ,i=1,2,...,n1],TR i is the actual fluctuation of electricity consumption on the i-th day, TR i The calculation method is:

[0030]

[0031] Among them, Max is the maximum value operation, || is the absolute value operation, is the daily minimum hourly electricity consumption on the i-th day, is the average daily electricity consumption on day i-1;

[0032] The value sequence of the decreasing trend of power consumption is DMM=[DMM i ,i=1,2,...,n1],DMM i is the downward trend value of power consumption on the i-th day, DMM i The calculation method is:

[0033]

[0034] in, The minimum hourly electricity consumption on the i-1th day;

[0035] The trend indicator sequence is DX=[DX i ,i=1,2,...,n1],DX i is the trend indicator of the i-th day, DX i The calculation method is:

[0036]

[0037] In one embodiment of the present invention, the BOLL indicator includes a middle line, an upper line, and a lower line.

[0038] The middle line is MD (P,n) , represents the standard deviation of daily electricity consumption in the n days before the i-th day, The calculation method is:

[0039]

[0040] Where n is the total number of days, P is the daily power consumption sequence, P = [P i, i=1,2,…,24],P i is the electricity consumption in the i-th hour of the day;

[0041] The upper rail is BOLL UP ,BOLL UP =[BOLL i UP ,i=1,2,…,n],BOLL i UP Indicates the upper line of the i-th day, BOLL i UP The calculation method is:

[0042] BOLL i UP =MA i (P,n) +α·MD i (P,n) ,

[0043] Among them, MA (P,n) Represents the moving average sequence of the previous n days of the daily electricity consumption sequence, MA (P,n) =[MA i (P,n) ,i=1,2,…,n],MA i (P,n) is the moving average of the electricity consumption on the i-th day in the previous n days, and α is the preset coefficient;

[0044] The lower rail is BOLL DN ,BOLL DN =[BOLL i DN ,i=1,2,…,n],BOLL i DN Indicates the lower line of the i-th day, BOLL i DN The calculation method is:

[0045] BOLL i DN =MA i (P,n) -α·MD i (P,n) .

[0046] The present invention also provides a financial indicator and neural network power consumption prediction system, comprising:

[0047] A data acquisition module is used to acquire the user's historical daily electricity consumption data, and extract the CCI index, DMI index, and BOLL index from the historical daily electricity consumption data according to the characteristics of financial indicators;

[0048] Model building module, used to build neural network models;

[0049] A data processing module, used to combine the historical daily power consumption data, CCI index, DMI index, BOLL index and divide them into a training set and a test set;

[0050] A training module, used to train the neural network model using the training set to obtain a trained neural network model;

[0051] The prediction module is used to input the test set into the trained neural network model to obtain the user's electricity consumption prediction result.

[0052] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method for predicting electricity consumption using financial indicators and a neural network is implemented.

[0053] The present invention also provides a device for predicting electricity consumption using financial indicators and neural networks, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method for predicting electricity consumption using financial indicators and neural networks when executing the computer program.

[0054] The above technical solution of the present invention has the following advantages compared with the prior art:

[0055] The present invention introduces the concept of technical indicators in the financial field, extracts features from the user's historical daily electricity consumption data according to the characteristics of financial indicators, and on this basis predicts the electricity consumption values ​​of power users in the future through a neural network; the present invention can effectively extract effective features from the user's historical electricity consumption data and improve the accuracy of the prediction of the user's electricity consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to make the content of the present invention more clearly understood, the present invention is further described in detail below according to specific embodiments of the present invention in conjunction with the accompanying drawings, wherein:

[0057] Figure 1 is a flow chart of the method of the present invention.

[0058] Figure 2 It is a flow chart of the present invention.

[0059] Figure 3 It is a corresponding diagram between the power consumption of power users and the CCI index in the embodiment of the present invention.

[0060] Figure 4 It is a corresponding diagram of the power consumption of the power user and the DMI index in the embodiment of the present invention.

[0061] Figure 5It is a corresponding diagram between the power consumption of power users and the BOLL index in the embodiment of the present invention.

[0062] Figure 6 It is a comparison chart of actual power consumption and predicted values ​​of different models in an embodiment of the present invention. DETAILED DESCRIPTION

[0063] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0064] Embodiment 1

[0065] Reference Figure 1 and Figure 2 As shown, the present invention discloses a method for predicting electricity consumption using financial indicators and neural networks, comprising the following steps:

[0066] S1: Obtain the user's historical daily electricity consumption data, and extract the CCI indicator, DMI indicator, and BOLL indicator from the historical daily electricity consumption data according to the characteristics of financial indicators. The CCI indicator, DMI indicator, and BOLL indicator can be used to extract the inherent trend characteristics of the electricity consumption data.

[0067] S1-1: Extract daily electricity consumption, daily maximum hourly electricity consumption, daily minimum hourly electricity consumption, daily average electricity consumption, average electricity consumption in the first 12 hours, and average electricity consumption in the last 12 hours from the user's historical daily electricity consumption data.

[0068] S1-2: The average daily electricity consumption is obtained based on the average electricity consumption in the first 12 hours and the average electricity consumption in the last 12 hours, and the CCI indicator, DMI indicator, and BOLL indicator are calculated in combination with other data in S1-1.

[0069] The CCI indicator is a CCI indicator value sequence, and the CCI indicator value sequence is CCI=[CCI i ,i=1,2,...,n],CCI i is the CCI indicator value of the i-th day in CCI, CCI i The calculation method is:

[0070]

[0071] Among them, TYP is the typical value sequence of power consumption, TYP = [TYP i ,i=1,2,...,n],TYP i is the typical value of electricity consumption on the ith day, n is the total number of days, and the typical value of electricity consumption is the average of the highest single-hour electricity consumption on that day, the lowest single-hour electricity consumption on that day, and the average daily electricity consumption on that day; MA (TYP,n)is the n-day simple moving average sequence, represents the simple moving average of the ith day in TYP; α is a preset coefficient, in this embodiment, α is 0.015, and n is 14. MAD represents the mean absolute error of TYP.

[0072] The MAD is the mean absolute error of TYP with a sliding window size of n days and a sliding step of 1 day. The calculation method of MAD is:

[0073]

[0074] The DMI index includes an electricity consumption increase index value sequence, an electricity consumption decrease index value sequence, and an average trend index sequence.

[0075] The power consumption increase index value sequence is PDI=[PDI i ,i=1,2,...,n1],PDI i is the power consumption increase index value on the i-th day, n1 is the preset number of days, PDI i The calculation method is:

[0076]

[0077] in, is the n1-day simple moving average of the rising trend value of electricity consumption on the i-th day, is the n1-day simple moving average of the actual fluctuation of electricity consumption on the i-th day;

[0078] The power consumption reduction index value sequence is MDI=[MDI i ,i=1,2,...,n1],MDI i is the electricity consumption reduction index value on the i-th day, MDI i The calculation method is:

[0079]

[0080] in, is the n1-day simple moving average of the downward trend value of electricity consumption on the i-th day;

[0081] The average trend index sequence is ADX=[ADX i ,i=1,2,...,n2],ADX i is the average trend indicator of the i-th day, n2 is the preset number of days, ADX i The calculation method is:

[0082]

[0083] in, is the simple moving average of the trend indicator sequence of the ith day over the n2 days. In this embodiment, n1=14, n2=6.

[0084] The value sequence of the rising trend of electricity consumption is DMP = [DMP i ,i=1,2,...,n1],DMP i is the rising trend value of electricity consumption on the i-th day, DMP i The calculation method is:

[0085]

[0086] in, is the highest single-hour electricity consumption on the ith day, is the maximum hourly electricity consumption on the i-1th day;

[0087] The real amplitude sequence of electricity consumption is TR = [TR i ,i=1,2,...,n1],TR i is the actual fluctuation of electricity consumption on the i-th day, TR i The calculation method is:

[0088]

[0089] Among them, Max is the maximum value operation, || is the absolute value operation, is the daily minimum hourly electricity consumption on the i-th day, is the average daily electricity consumption on day i-1;

[0090] The value sequence of the decreasing trend of power consumption is DMM=[DMM i ,i=1,2,...,n1],DMM i is the downward trend value of power consumption on the i-th day, DMM i The calculation method is:

[0091]

[0092] in, The minimum hourly electricity consumption on the i-1th day;

[0093] The trend indicator sequence is DX=[DX i ,i=1,2,...,n1],DX i is the trend indicator of the i-th day, DX i The calculation method is:

[0094]

[0095] The BOLL indicator includes the middle line, the upper line and the lower line.

[0096] The middle track is It represents the standard deviation of daily electricity consumption in the n days before the i-th day, which is used to grasp the overall trend of the electricity consumption curve. The calculation method is:

[0097]

[0098] Where n is the total number of days, P is the daily power consumption sequence, P = [P i , i=1,2,…,24],P i is the electricity consumption in the i-th hour of the day;

[0099] The upper rail is BOLL UP ,BOLL UP =[BOLL i UP ,i=1,2,…,n],BOLL i UP Indicates the upper line of the i-th day, BOLL i UP The calculation method is:

[0100]

[0101] Among them, MA (P,n) Represents the moving average sequence of the previous n days of the daily electricity consumption sequence, MA (P,n) =[MA i (P,n) ,i=1,2,…,n],MA i (P,n) is the moving average of the electricity consumption on the ith day in the previous n days, and α is the preset coefficient. The larger the value of α is, the larger the bandwidth of the BOLL indicator is, and the better the tolerance to fluctuations in the electricity consumption curve is.

[0102] The lower rail is BOLL DN ,BOLL DN =[BOLL i DN ,i=1,2,…,n],BOLL i DN Indicates the lower line of the i-th day, BOLL i DN The calculation method is:

[0103] BOLL i DN =MA i (P,n) -α·MD i (P,n) .

[0104] BOLL UP and BOLL DN Define upper and lower bounds for fluctuations in the electricity consumption curve.

[0105] S2: The historical daily electricity consumption data, CCI index, DMI index, and BOLL index are combined and normalized, and the normalized data is divided into a training set and a test set.

[0106] S2-1: The data set is obtained by combining daily electricity consumption, daily maximum hourly electricity consumption, daily minimum hourly electricity consumption, average electricity consumption in the first 12 hours, average electricity consumption in the last 12 hours, CCI index value sequence, electricity consumption increase index value sequence, electricity consumption decrease index value sequence, average trend index sequence, upper rail line, middle rail line, lower rail line and period parameters. The specific composition of the data set is shown in Table 1.

[0107] Table 1 Specific composition of the data set

[0108]

[0109] The data set constructed in this embodiment includes 13 dimensions of data, including a 5-dimensional power consumption input vector, a 7-dimensional financial technical indicator input vector, and a 1-dimensional social factor indicator. The period parameter Day represents Monday to Sunday, and the quantized value is Day = {1, 2, 3, 4, 5, 6, 7}.

[0110] S2-2: Normalize the data set obtained in S2-1 and divide the normalized data into a training set and a test set.

[0111] S3: construct a neural network model, use the training set to train the neural network model to obtain a trained neural network model, and input the test set into the trained neural network model to obtain a user power consumption prediction result.

[0112] S3-1: The neural network model includes a convolutional neural network (CNN) and a bidirectional long short-term memory neural network (Bi-directional Long Short-Term Memory, BiLSTM).

[0113] The convolutional neural network extracts local features of the training set and the test set. The convolutional neural network in this embodiment is a one-dimensional convolutional neural network (1DCNN), which includes a one-dimensional convolutional layer and a maximum pooling layer. The one-dimensional convolutional layer is used to extract effective nonlinear local features, and the maximum pooling layer is used to compress the extracted effective feature information and obtain more representative information.

[0114] The bidirectional long short-term memory neural network is trained using local features extracted from the training set by the convolutional neural network, and the local features extracted from the prediction set by the convolutional neural network are input into the trained bidirectional long short-term memory neural network to obtain user electricity consumption prediction data.

[0115] S3-2: Denormalize the user's electricity consumption forecast data to obtain the user's daily electricity consumption forecast result.

[0116] Embodiment 2

[0117] The present invention also discloses a financial indicator and neural network power consumption prediction system, comprising:

[0118] A data acquisition module is used to acquire the user's historical daily electricity consumption data, and extract the CCI index, DMI index, and BOLL index from the historical daily electricity consumption data according to the characteristics of financial indicators;

[0119] Model building module, used to build neural network models;

[0120] A data processing module, used to combine the historical daily power consumption data, CCI index, DMI index, BOLL index and divide them into a training set and a test set;

[0121] A training module, used to train the neural network model using the training set to obtain a trained neural network model;

[0122] The prediction module is used to input the test set into the trained neural network model to obtain the user's electricity consumption prediction result.

[0123] Embodiment 3

[0124] The present invention also discloses a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method for predicting electricity consumption using financial indicators and a neural network in the first embodiment is implemented.

[0125] Embodiment 4

[0126] The present invention also discloses a device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for predicting electricity consumption using financial indicators and neural networks in Example 1 is implemented.

[0127] The present invention introduces the concept of technical indicators in the financial field, extracts features from the user's historical daily electricity consumption data according to the characteristics of financial indicators, and on this basis predicts the electricity consumption values ​​of power users in the future through a neural network; the present invention can effectively extract effective features from the user's historical electricity consumption data and improve the accuracy of the prediction of the user's electricity consumption.

[0128] In order to further demonstrate the beneficial effects of the present invention, in this embodiment, the method of the present invention (represented by CNN-BiLSTM-FI) and other methods in the prior art are used to predict user electricity consumption.

[0129] Using the feature extraction method of the present invention, the CCI index extracted from the power consumption data of power users is as follows: Figure 3 As shown in Figure 2, the DMI index extracted from the power consumption data of power users is as follows: Figure 4 As shown in the figure, the BOLL indicators extracted from the electricity consumption data of power users are as follows: Figure 5 As shown, Figure 5 The arrows indicate the portion of electricity consumption that exceeds the upper and lower lines.

[0130] First, when the feature extraction method according to the characteristics of financial indicators in the present invention is not used, the user's electricity consumption is predicted, and the prediction results of different models are shown in Table 2; then, when the feature extraction method according to the characteristics of financial indicators in the present invention is used, the user's electricity consumption is predicted, and the prediction results of different models are shown in Table 3. The prediction effect is reflected by the mean absolute error (MAE) and the root mean square error (RMSE). The smaller the values ​​of MAE and RMSE, the higher the prediction accuracy.

[0131] Table 2 Prediction results of each model when the feature extraction method in the present invention is not used

[0132] Model MAE RMSE LSTM 30.957 35.974 GRU 34.156 39.506 BiLSTM 28.058 31.761 CNN-LSTM 21.574 25.521 CNN-BiLSTM 14.412 19.852

[0133] Table 3 Prediction results of each model when using the feature extraction method of the present invention

[0134] Model MAE RMSE LSTM-FI 26.198 31.152 GRU-FI 28.697 33.452 BiLSTM-FI 21.502 24.843 CNN-LSTM-FI 16.684 20.782 CNN-BiLSTM-FI 13.227 16.673

[0135] In Tables 2 and 3, LSTM represents a long short-term memory network, GRU represents a recurrent gate unit in a recurrent neural network, BiLSTM represents a bidirectional long short-term memory neural network, CNN-LSTM represents a combination of a convolutional neural network and a long short-term memory network, CNN-BiLSTM represents a method of the present invention that does not extract features based on the characteristics of financial indicators, and FI represents the method of extracting features based on the characteristics of financial indicators in the present invention. The parameters of each model are set as follows: the number of convolutional layers is 1, the number of convolutional kernels is 64, the number of training epochs is 300, the learning rate is 0.001, the batch training size is 16, and the forgetting rate is 0.2.

[0136] It can be seen from Table 2 and Table 3 that the prediction effect of the present invention is the best when the feature extraction according to the characteristics of the financial indicators in the present invention is not used and when the feature extraction according to the characteristics of the financial indicators in the present invention is used.

[0137] Finally, the power consumption predicted by each method is compared with the actual power consumption. The results are as follows: Figure 6 As shown, from Figure 6 It can be seen that the prediction accuracy of the method of the present invention is also the highest.

[0138] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0139] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0140] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0141] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0142] Obviously, the above embodiments are merely examples for clear explanation and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived from these are still within the protection scope of the invention.

Claims

1. A method for predicting electricity consumption using financial indicators and neural networks, characterized in that: include: Obtain the user's historical daily electricity consumption data, extract the CCI index, the DMI index, and the BOLL index from the historical daily electricity consumption data according to the characteristics of the financial indicators, combine the historical daily electricity consumption data, the CCI index, the DMI index, and the BOLL index, and divide them into a training set and a test set; A neural network model is constructed, the training set is used to train the neural network model to obtain a trained neural network model, and the test set is input into the trained neural network model to obtain a user power consumption prediction result.

2. The power consumption prediction method based on financial indicators and neural networks according to claim 1 is characterized by: The neural network model includes a convolutional neural network and a bidirectional long short-term memory neural network. The convolutional neural network extracts local features of the training set and the test set, uses the local features extracted by the convolutional neural network from the training set to train the bidirectional long short-term memory neural network, and inputs the local features extracted by the convolutional neural network from the prediction set into the trained bidirectional long short-term memory neural network to obtain the user power consumption prediction result.

3. The electricity consumption prediction method based on financial indicators and neural networks according to claim 1 is characterized by: The CCI indicator is a CCI indicator value sequence, and the CCI indicator value sequence is CCI=[CCI i ,i=1,2,...,n],CCI i is the CCI indicator value of the i-th day in CCI, CCI i The calculation method is: Among them, TYP is the typical value sequence of power consumption, TYP = [TYP i ,i=1,2,...,n],TYP i is the typical value of electricity consumption on the ith day, n is the total number of days, and the typical value of electricity consumption is the average of the highest single-hour electricity consumption, the lowest single-hour electricity consumption, and the average daily electricity consumption of the day; MA (TYP,n) is the n-day simple moving average sequence, Represents the simple moving average of the i-th day in TYP; α is the preset coefficient, and MAD represents the mean absolute error of TYP.

4. The method for predicting electricity consumption using financial indicators and neural networks according to claim 3 is characterized by: The MAD is the mean absolute error of TYP with a sliding window size of n days and a sliding step of 1 day. The calculation method of MAD is:

5. The method for predicting electricity consumption using financial indicators and neural networks according to claim 1 is characterized by: The DMI index includes an electricity consumption increase index value sequence, an electricity consumption decrease index value sequence, and an average trend index sequence. The power consumption increase index value sequence is PDI=[PDI i ,i=1,2,...,n1],PDI i is the power consumption increase index value on the i-th day, n1 is the preset number of days, PDI i The calculation method is: in, is the n1-day simple moving average of the rising trend value of electricity consumption on the i-th day, is the n1-day simple moving average of the actual fluctuation of electricity consumption on the i-th day; The power consumption reduction index value sequence is MDI=[MDI i ,i=1,2,...,n1],MDI i is the electricity consumption reduction index value on the i-th day, MDI i The calculation method is: in, is the n1-day simple moving average of the downward trend value of electricity consumption on the i-th day; The average trend index sequence is ADX=[ADX i ,i=1,2,...,n2],ADX i is the average trend indicator of the i-th day, n2 is the preset number of days, ADX i The calculation method is: in, It is the n2-day simple moving average of the trend indicator sequence on the i-th day.

6. The electricity consumption prediction method based on financial indicators and neural networks according to claim 5 is characterized by: The value sequence of the rising trend of electricity consumption is DMP = [DMP i ,i=1,2,...,n1],DMP i is the rising trend value of electricity consumption on the i-th day, DMP i The calculation method is: in, is the highest single-hour electricity consumption on the ith day, is the maximum hourly electricity consumption on the i-1th day; The real amplitude sequence of electricity consumption is TR = [TR i ,i=1,2,...,n1],TR i is the actual fluctuation of electricity consumption on the i-th day, TR i The calculation method is: Among them, Max is the maximum value operation, || is the absolute value operation, is the daily minimum hourly electricity consumption on the i-th day, is the average daily electricity consumption on day i-1; The value sequence of the decreasing trend of power consumption is DMM=[DMM i ,i=1,2,...,n1],DMM i is the downward trend value of power consumption on the i-th day, DMM i The calculation method is: in, The minimum hourly electricity consumption on the i-1th day; The trend indicator sequence is DX=[DX i ,i=1,2,...,n1],DX i is the trend indicator of the i-th day, DX i The calculation method is:

7. The method for predicting electricity consumption using financial indicators and neural networks according to any one of claims 1 to 6, characterized in that: The BOLL indicator includes the middle line, the upper line and the lower line. The middle line is MD (P,n) , represents the standard deviation of daily electricity consumption in the n days before the i-th day, The calculation method is: Where n is the total number of days, P is the daily power consumption sequence, P = [P i , i=1,2,…,24],P i is the electricity consumption in the i-th hour of the day; The upper rail is BOLL UP , Indicates the upper line of the i-th day, BOLL i UP The calculation method is: BOLL i UP =MA i (P,n) +α·MD i (P,n) , Among them, MA (P,n) Represents the moving average sequence of the previous n days of the daily electricity consumption sequence, MA (P,n) =[MA i (P,n) ,i=1,2,…,n],MA i (P,n) is the moving average of the electricity consumption on the i-th day in the previous n days, and α is the preset coefficient; The lower rail is BOLL DN ,BOLL DN =[BOLL i DN ,i=1,2,…,n],BOLL i DN Indicates the lower line of the i-th day, BOLL i DN The calculation method is: BOLL i DN =MA i (P,n) -α·MD i (P,n) .

8. A power consumption prediction system based on financial indicators and neural networks, characterized in that: include: A data acquisition module is used to acquire the user's historical daily electricity consumption data, and extract the CCI index, DMI index, and BOLL index from the historical daily electricity consumption data according to the characteristics of financial indicators; Model building module, used to build neural network models; A data processing module, used to combine the historical daily power consumption data, CCI index, DMI index, BOLL index and divide them into a training set and a test set; A training module, used to train the neural network model using the training set to obtain a trained neural network model; The prediction module is used to input the test set into the trained neural network model to obtain the user's electricity consumption prediction result.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting electricity consumption using a financial indicator and a neural network as described in any one of claims 1 to 7 is implemented.

10. A device for predicting electricity consumption based on financial indicators and neural networks, characterized in that: The method comprises 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 electricity consumption using a financial indicator and a neural network as described in any one of claims 1 to 7 is implemented.

Images