Information processing device, information processing method, and information processing program

Abstract

To provide an information processing device capable of predicting power market price with high accuracy.SOLUTION: An information processing device to predict power price at a predicting time includes at least the spare rate based on the past electricity price, a power demand, and power supply capability. The information processing device includes an acquisition unit to acquire the past data including the related information related to power price, an extraction unit to extract the similarity data that the related information of the past data, of the past data, is similar to the related information at the prediction time, a correction unit to correct the power price of the similarity data based on the difference between the spare rate included in the related information of the similarity data and the spare rate at the prediction time, and a prediction unit to predict the power price at the prediction time based on the corrected power price.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an information processing device, an information processing method, and an information processing program. [Background technology]

[0002] The price of electricity traded in the electricity trading market (hereinafter referred to as "electricity market price") fluctuates due to various highly nonlinear factors, making it difficult to predict. In response to this, electricity market price prediction techniques based on various input data have been proposed (see, for example, Patent Documents 1 and 2). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2019-046281 [Patent Document 2] Japanese Patent Application Publication No. 2020-067919 Summary of the Invention [Problem to be solved by the invention]

[0004] Electricity market prices are generally determined by the balance of supply and demand (supply against demand), but the tendency for electricity market prices to change differs depending on whether or not the supply and demand is tight. Therefore, technology that does not take into account tight supply and demand of electricity cannot respond to fluctuations in electricity market prices when the supply and demand is tight, making it difficult to accurately predict electricity market prices.

[0005] The present invention has been made in view of the above-mentioned problems, and has an object to provide an information processing device, an information processing method, and an information processing program that are capable of predicting electricity market prices with high accuracy. [Means for solving the problem]

[0006] One invention for achieving the above object is an information processing device that predicts electricity prices at a prediction time point, the information processing device including: an acquisition unit that acquires past data including past electricity prices and related information related to electricity prices including at least a reserve margin based on electricity demand and electricity supply capacity; an extraction unit that extracts similar data from the past data, the related information of which is similar to the related information at the prediction time point; a correction unit that corrects the electricity price of the similar data based on the difference between the reserve margin included in the related information of the similar data and the reserve margin at the prediction time point; and a prediction unit that predicts the electricity price at the prediction time point based on the corrected electricity price.

[0007] Also, an information processing method in which an information processing device predicts an electricity price at a prediction time includes the steps of: acquiring historical data including past electricity prices and related information related to electricity prices, including at least a reserve margin based on electricity demand and electricity supply capacity; extracting similar data from the historical data, whose related information is similar to the related information at the prediction time; correcting the electricity price of the similar data based on the difference between the reserve margin included in the related information of the similar data and the reserve margin at the prediction time; and predicting the electricity price at the prediction time based on the corrected electricity price.

[0008] The present invention also provides an information processing program for predicting an electricity price at a time of prediction, the information processing program causing a computer to implement the following: an acquisition unit that acquires past data including past electricity prices and related information related to electricity prices including at least a reserve margin based on electricity demand and electricity supply capacity; an extraction unit that extracts similar data from the past data, the related information of which is similar to the related information at the time of prediction; a correction unit that corrects the electricity price of the similar data based on a difference between the reserve margin included in the related information of the similar data and the reserve margin at the time of prediction; and a prediction unit that predicts the electricity price at the time of prediction based on the corrected electricity price. Other features of the present invention will become clear from the description in this specification. [Effects of the Invention]

[0009] According to the present invention, it is possible to provide an information processing device that can predict the electricity market price with high accuracy. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 2 is a configuration diagram illustrating an example of functions and databases included in an information processing device. [Figure 2] FIG. 10 is a diagram illustrating an outline of a process executed by a reserve rate information management unit. [Figure 3] FIG. 2 is a diagram illustrating an example of detailed functional blocks of a prediction model construction and inference unit. [Figure 4] 10 is a flowchart illustrating an electricity market price prediction process executed by an information processing device. [Figure 5] FIG. 10 is a diagram illustrating an example of request point data Xquery and learning data Ylearning and Xlearning. [Figure 6] FIG. 2 is a diagram illustrating a hardware configuration of an information processing device. DETAILED DESCRIPTION OF THE INVENTION

[0011] The environment surrounding electric power companies in Japan has become increasingly competitive since the unbundling of power generation and transmission in 2015. As a result of this separation, electric power companies have been separated into freely competitive power generation companies, retailers, and neutral transmission and distribution companies, placing importance on trading in the electricity market to achieve economic efficiency and grid stability. Electricity markets include the Japan Electric Power Exchange (JEPX), a market where economically rational amounts of electricity are traded to achieve simultaneous balancing of planned values ​​between retailers and power generation companies, and the Electric Power Reserve Exchange (EPRX), a market where transmission and distribution companies trade adjustment capacity to achieve grid stability after gate closure. Retailers and power generation companies need to predict these electricity market prices in advance in order to increase their profits from electricity trading.

[0012] <<Information processing device 10>> 1 is a configuration diagram showing an example of functions and databases provided in an information processing device 10. The information processing device 10 is an information processing device that predicts the electricity market price, which is the price of electricity traded in the electricity trading market. The information processing device 10 is configured to include the following functional units: an acquisition unit 11, a database 12, a reserve margin information management unit 13, and a prediction model construction and inference unit 14.

[0013] Each of these will be explained below, but first an overview will be given here, and the details of each will be explained later with specific examples using flowcharts.

[0014] [Acquisition part 11] The acquisition unit 11 is a functional unit that acquires past data including actual values ​​of past electricity market prices and related information related to the electricity market prices. For example, the acquisition unit 11 acquires information published on the webpage of a public institution such as the Organization for Cross-regional Coordination of Transmission Operators, JAPAN (OCCTO). The acquisition unit 11 also acquires related information for the target prediction date. The target prediction date is an example of the prediction time point. The acquisition unit 11 writes the acquired information into the database 12. The acquisition unit 11 may accept input from an input device such as a mouse, keyboard, or touch panel, or may receive information from another device via a wired or wireless communication network such as the Internet, a local area network (LAN), a wide area network (WAN), or a dedicated line.

[0015] The related information includes at least reserve margin information and is information that may be a factor in fluctuations in the electricity market price. The reserve margin information is data indicating the value of the reserve margin based on the demand for electricity and the supply capacity of electricity. The related information includes, for example, actual values, forecast values, planned values, or calculated values ​​of electricity market information related to the electricity market, supply and demand information related to the supply and demand of electricity, interconnection line information related to interconnection lines, generator information related to generators, fuel information, meteorological information related to weather, calendar information related to calendars, and reserve margin information. Examples of the electricity market information include, for example, market price, contracted amount, and bid amount. Examples of the supply and demand information include, for example, demand and supply capacity. Examples of the interconnection line information include, for example, available capacity. Examples of the generator information include, for example, operation amount, output suspension amount, output reduction amount, and the like. Examples of the fuel information include, for example, crude oil price and natural gas price. Examples of the weather information include, for example, temperature and humidity. Examples of the calendar information include, for example, public holidays, weekdays, and the like. Note that the related information may be values ​​for each area or each location.

[0016] The acquisition unit 11 may acquire reserve margin information published by a public institution, or may calculate the reserve margin using the following formula (1) based on supply and demand information included in the related information.

[0017] Reserve margin = supply capacity ÷ demand × 100 (1)

[0018] In other words, the reserve margin is the ratio of electricity supply capacity to electricity demand, and is an indicator of whether a power plant's "supply capacity" is sufficient to meet the "electricity demand."

[0019] [Database 12] The database 12 stores information (for example, a history of past data and information related to the target prediction date) acquired by the acquisition unit 11. The database 12 may also store other data or intermediate results of calculations in each functional unit.

[0020] [Reserve Margin Information Management Department 13] The reserve margin information management unit 13 is a functional unit that sets a plurality of reserve margin intervals with different trends in changes in the electricity market price. The reserve margin information management unit 13 corresponds to a "setting unit."

[0021] FIG. 2 is a diagram outlining the processing executed by the reserve margin information management unit 13. The vertical axis of graphs G1 and G2 shown in this figure represents the past actual values ​​of the electricity market price, and the horizontal axis represents the past actual values ​​of the reserve margin. That is, graph G1 represents the correlation between the electricity market price and the reserve margin. As shown in the figure, there is a negative correlation between the electricity market price and the reserve margin. That is, the larger the reserve margin, the lower the electricity market price, and the smaller the reserve margin, the higher the electricity market price. Furthermore, the electricity market price is characterized by a steeper slope during times of tight supply and demand when the reserve margin becomes smaller (e.g., interval A in graph G2). The reserve margin information management unit 13 may automatically set multiple reserve margin intervals using a clustering technique or a decision tree technique based on the relationship between the past actual values ​​of the electricity market price and the past actual values ​​of the reserve margin. In the example shown in graph G2, the reserve margin information management unit 13 sets two reserve margin intervals, interval A and interval B. It should be noted that the present invention is not limited to this example, and the reserve rate information management unit 13 may set three or more reserve rate intervals.

[0022] For example, when a clustering method is used, the reserve margin information management unit 13 sets multiple intervals of the clustered reserve margin by applying VGMM (Gaussian Mixture Model) that divides data indicating multiple reserve margins based on the distribution shape of the electricity market price, or DBSCAN (Density-Based Spatial Clustering of Applications with Noise) that divides based on density, etc. When a decision tree is used, the reserve margin information management unit 13 sets multiple intervals of the reserve margin based on the tendency of changes in the electricity market price relative to the reserve margin by using the electricity market price as the objective variable and the reserve margin as the explanatory variable.

[0023] The reserve rate information management unit 13 may receive the setting input of the section from the user, for example, by displaying the graph G1.

[0024] [Predictive Model Building and Inference Section 14] The prediction model construction and inference unit 14 constructs a prediction model for predicting the electricity market price, and predicts the electricity market price on the target prediction date using the constructed prediction model. Specifically, the prediction model construction and inference unit 14 constructs a prediction model in which the electricity market price in the interval corresponding to the reserve margin on the target prediction date is used as the objective variable and related information is used as the explanatory variables. For example, if the reserve margin on the target prediction date is included in interval A, the prediction model construction and inference unit 14 extracts from the database 12 past data for the date and time when the reserve margin falls within interval A, and constructs a prediction model using the extracted past data as learning data. In other words, the prediction model construction and inference unit 14 constructs a different prediction model for each interval.

[0025] The prediction model construction and inference unit 14 may construct a prediction model using a different method for each interval. For example, the prediction model construction and inference unit 14 may apply a physical model such as a fundamental model, a statistical model such as a multiple regression equation, JIT (Just In Time) modeling, SVR (Support Vector Regression), or a machine learning model using learning. The prediction model construction and inference unit 14 then inputs information related to the prediction target date as explanatory variables into the constructed prediction model and obtains the output, thereby predicting the electricity market price on the prediction target date.

[0026] 3 is a diagram showing an example of detailed functional blocks of the prediction model construction / inference unit 14. This diagram shows an example of a configuration when forecasting electricity market prices using a prediction model constructed by JIT modeling. The prediction model construction / inference unit 14 is configured to include the following functional units: an extraction unit 141, a correction unit 142, a prediction unit 143, and an output unit 144.

[0027] [Extraction part 141] The extraction unit 141 extracts similar data, the related information of which is similar to the related information at the time of prediction, from the past data stored in the database 12. The similar data corresponds to the neighboring data described later. For example, the extraction unit 141 extracts the request point data X based on the past data stored in the database 12. query , and training data Y learning ,X learning Generate the required point data X query , and training data Y learning ,X learning Based on this, similar data is extracted.

[0028] [Correction unit 142] The correction unit 142 corrects the electricity market price of the similar data based on the difference between the reserve margin included in the related information of the similar data and the reserve margin at the time of prediction.

[0029] [Prediction Section 143] The prediction unit 143 predicts the electricity market price at the time of prediction based on the corrected electricity market price.

[0030] [Output section 144] The output unit 144 outputs the predicted electricity market price for the prediction target date. For example, the output unit 144 may display the predicted electricity market price on a display device, may transmit it to another device via a network, or may output it using a printer or the like.

[0031] <<Processing Executed by Information Processing Device 10>> The process executed by the information processing device 10 will be described using a flowchart. Fig. 4 is a flowchart illustrating the electricity market price prediction process executed by the information processing device 10. In this embodiment, an example will be described in which the electricity market price is predicted using a prediction model constructed by JIT modeling.

[0032] First, in step S101, the extraction unit 141 extracts from the database 12 past data for the interval that includes the reserve margin for the target prediction date. The past data for the interval that includes the reserve margin for the target prediction date is related information at the time point when the reserve margin falls within the relevant interval and actual values ​​of past electricity market prices. For example, if the target prediction date is "2022 / 1 / 1 0:00" and the reserve margin for the target prediction date is "30%", the extraction unit 141 extracts past data for the time point when the reserve margin falls within the range of interval A (e.g., 0% to 40%) shown in graph G2 of FIG. 2 during the target period from 2021 / 1 / 1 0:00 to 2021 / 12 / 31 0:00. Then, based on the extracted past data, the extraction unit 141 extracts request point data X query , and training data Y learning ,X learning Generate.

[0033] <Required point data and learning data> Figure 5 shows the required point data X query , and training data Y learning ,X learning This figure shows an example of the electricity market price Y query This shows an example of data when predicting the required point data X query is an explanatory variable for the prediction target date "2022 / 1 / 1 0:00". The explanatory variables are fluctuation factors of the electricity market price, including at least the reserve margin, among the related information acquired by the acquisition unit 11. By including the reserve margin in the explanatory variables as a fluctuation factor of the electricity market price, it is possible to take into account the tight demand situation. In the example shown in the figure, the day-ahead market price, the current day demand, the reserve margin, and the current day weekday / holiday flag are used as explanatory variables, but this is not limited to these, and other information included in the related information may also be used as explanatory variables. The day-ahead market price is the electricity market price on the day before the prediction target date. The current day demand is the demand on the prediction target date. The reserve margin is the reserve margin on the prediction target date. The current day weekday / holiday flag is a flag that indicates whether the prediction target date is a weekday or a holiday.

[0034] Training data Y learning ,X learningis the objective variable Y , which is the electricity market price for a given period in the past (in the illustrated example, one year from 2021 / 1 / 1 0:00 to 2021 / 12 / 31 0:00). learning , and explanatory variables X learning The training data Y learning ,X learning is past data for a time point where the reserve margin is within interval A (e.g., 0% to 40%), which corresponds to the reserve margin of 30% at the prediction time point "2022 / 1 / 1 0:00" (i.e., the reserve margin is 0% to 40%). Here, interval A is, for example, interval A shown in graph G2 of FIG. 2. Note that while this example illustrates a case where the objective variable is daily, the present invention is not limited to this, and electricity market prices for each hour or time period, etc., may be used in accordance with the prediction target.

[0035] Subsequently, in step S102, the extraction unit 141 extracts the requested point data X query Training data X for learning For example, the extraction unit 141 calculates the distance measure using a distance function such as a weighted distance as shown in the following formula (1). At this time, the weight α(i) uses a value calculated from the learning data using a nonlinear variable selection method such as the MIC (Maximal Information Coefficient) of each explanatory variable for the objective variable or variable importance. Furthermore, a normalized or standardized value is used for each explanatory variable.

[0036]

number

[0037] where D(n) is the nth training data X learning is the distance measure of the i-th explanatory variable, S is the number of explanatory variables, α(i) is the weight of the i-th explanatory variable, and X query (i) is the required point data X query is the value of the i-th explanatory variable in X learning (n,i) is the nth training data X learning That is, the extraction unit 141 extracts the value of the i-th explanatory variable X query and explanatory variables of the training data Xlearning Based on the difference between learning Each distance measure is calculated.

[0038] Next, in step S103, the extraction unit 141 extracts an arbitrary number K (K is a positive integer) of learning data in ascending order of the distance measure as neighboring data. Here, extracting similar data means extracting a predetermined number (K) of data with a small distance measure. Note that the value of K may be 1. As a result, the extraction unit 141 extracts the requested point data X query Therefore, even if the reserve margin for the target forecast date has not been experienced in the past, appropriate nearby data can be generated.

[0039] Next, in step S104, the correction unit 142 corrects the neighborhood data based on the reserve rate. Since the extracted neighborhood data is past data, there is a discrepancy between the reserve rate for the forecast target date and the reserve rate for the neighborhood data, and this difference becomes an error. This is also true for other explanatory variables other than the reserve rate. Therefore, the correction unit 142 uses a correction model h( X query , X neighbor , Y neighbor ) to calculate the target variable Y neighbor Correct (k).

[0040]

number

[0041] where a(i) is the regression coefficient of the i-th explanatory variable. For example, the correction unit 142 calculates the regression coefficient a(i) of each explanatory variable from learning data in which the electricity market price for an arbitrary past period is the objective variable and related information is the explanatory variable. neighbor (k) is the target variable of the kth corrected neighbor data, and X neighbor are explanatory variables for the neighborhood data, and Y neighbor is the target variable of the neighborhood data, and S^ is the number of explanatory variables. neighbor(k,i) is the i-th explanatory variable of the k-th neighbor data, and Y neighbor (k) is the objective variable of the k-th neighborhood data. That is, the correction unit 142 calculates the explanatory variable X query and the explanatory variables of the neighborhood data X neighbor Based on the difference between neighbor Correct the following.

[0042] Although the above-mentioned equation (2) is a linear equation, a correction model such as a quadratic equation may be constructed.

[0043] Subsequently, in step S105, the prediction unit 143 predicts the electricity market price by integrating the corrected neighborhood data. For example, the prediction unit 143 predicts the electricity market price by using a statistical model (prediction model) f(Ŷ neighbor ) to predict the electricity market price Y. That is, the prediction unit 143 uses a statistic such as the average of the objective variable (electricity market price) of the corrected neighborhood data as the predicted value Y of the electricity market price.

number

[0044] Here, Y is the predicted value of the electricity market price, and K is the number of neighboring data.

[0045] In the above-described formula (3), the average value of the electricity market prices of the corrected neighborhood data is set as the predicted value Y of the electricity market price, but this is not limiting. For example, the prediction unit 143 may calculate the predicted value of the electricity market price by weighting each neighborhood data by a distance measure. In this case, the statistical value of the neighborhood data weighted by the distance measure corresponds to the statistical quantity. In this case, the weight of each neighborhood data may be the reciprocal of the distance measure.

[0046] Subsequently, in step S106, the output unit 144 outputs the predicted value Y of the electricity market price, and the electricity market price prediction process ends.

[0047] The process of setting the reserve margin interval by the reserve margin information management unit 13 only needs to be executed at least once before the electricity market price prediction process in which the prediction model construction / inference unit 14 predicts the electricity market price, and may be executed at a cycle different from the electricity market price prediction process. For example, the information processing device 10 may set the reserve margin interval every time the electricity market price is predicted, or may set the reserve margin interval at a cycle (e.g., monthly) longer than the cycle (e.g., daily) in which the electricity market price is predicted.

[0048] Furthermore, the information processing device 10 may perform the electricity market price forecasting process without setting a reserve margin interval. When performing the electricity market price forecasting process without setting a reserve margin interval, the information processing device 10 extracts all past data regardless of the reserve margin classification in the above-mentioned S101. Furthermore, the information processing device 10 may omit the process of the above-mentioned S104.

[0049] <Hardware configuration of information processing device 10> 6 is a diagram illustrating the hardware configuration of the information processing device 10. The information processing device 10 is a computer having a CPU (Central Processing Unit) 101, a memory 102, an input unit 103, an output unit 104, a storage unit 105, a recording medium drive unit 106, and a network connection unit 107.

[0050] [CPU101] The CPU 101 executes information processing programs stored in the memory 102 or the storage unit 105 to realize various functions of the information processing device 10 (for example, the functions of the acquisition unit 11, the reserve rate information management unit 13, and the prediction model construction / inference unit 14).

[0051] [Memory 102] The memory 102 is, for example, a RAM (Random-Access Memory) and is used as a temporary storage area for various programs, data, and the like.

[0052] [Input section 103] The input unit 103 is a device that accepts commands and data input by the user, and includes an input interface such as a keyboard and a touch sensor that detects a touch position on a touch panel display.

[0053] [Output section 104] The output unit 104 is, for example, a device such as a display or a printer.

[0054] [Storage section 105] The storage unit 105 is a non-transitory (for example, non-volatile) storage device that stores various data to be executed or processed by the CPU 101. The storage unit 105 stores the database 12, for example.

[0055] [Recording medium drive unit 106] The recording medium driving unit 106 reads various data such as information processing programs recorded on a recording medium such as an SD card, DVD, or CD-ROM, and stores the data in the storage unit 105 .

[0056] [Network Connection Section 107] A network connection unit 107 exchanges various programs and data with other computers via a network.

[0057] =====Summary===== As described above, the information processing device 10 of the embodiment is an information processing device 10 that predicts the electricity market price at the time of prediction, and includes an acquisition unit 11 that acquires past data including past electricity market prices and related information related to the electricity market price including at least a reserve margin based on electricity demand and electricity supply capacity, an extraction unit 141 that extracts similar data (neighboring data) from the past data whose related information is similar to the related information at the time of prediction, a correction unit 142 that corrects the electricity market price of the similar data based on the difference between the reserve margin included in the related information of the similar data and the reserve margin at the time of prediction, and a prediction unit 143 that predicts the electricity market price at the time of prediction based on the corrected electricity market price.

[0058] With this configuration, it is possible to predict the electricity market price with high accuracy.

[0059] The information processing device 10 further includes a reserve margin information management unit 13 that sets multiple reserve margin intervals with different trends in changes in the electricity market price based on the electricity price and reserve margin included in the past data, and the extraction unit 141 extracts similar data from the past data of the reserve margin interval at the time of prediction.

[0060] According to this configuration, it is possible to predict the electricity market price by excluding past data from other intervals where the trend of change in the electricity market price is different, thereby further improving the prediction accuracy. For example, when the reserve margin at the time of prediction is near the border between a state where the supply and demand of electricity is tight (e.g., interval A shown in FIG. 2) and a state where it is not (e.g., interval B shown in FIG. 2), if similar data is extracted from all past data, the prediction accuracy may decrease. In contrast, by excluding past data from other intervals where the trend of change in the electricity market price is different (e.g., interval B when the reserve margin at the time of prediction is in interval A) (i.e., using only past data from interval A), it is possible to make a prediction that corresponds to the trend of change in the electricity market price.

[0061] In the information processing device 10, the reserve rate information management unit 13 sets a plurality of reserve rate intervals by clustering or a decision tree.

[0062] This configuration eliminates the need for users to manually set the categories, and also allows for accurate setting of reserve margin intervals that reflect different trends in electricity market price changes, thereby further improving forecast accuracy.

[0063] In addition, in the information processing device 10, the extraction unit 141 extracts multiple similar data, the correction unit 142 corrects the electricity market price of each of the multiple similar data, and the prediction unit 143 predicts the electricity price on the predicted date based on statistics of the multiple corrected electricity market prices.

[0064] With this configuration, it is possible to further improve the accuracy of forecasting the electricity market price.

[0065] Furthermore, the information processing device 10 of the embodiment is an information processing device 10 that predicts the electricity market price at the time of prediction, and includes: an acquisition unit 11 that acquires past data including past electricity market prices and related information related to the electricity market price including at least a reserve margin based on electricity demand and electricity supply capacity; a reserve margin information management unit 13 that sets multiple reserve margin ranges with different trends in electricity market price change based on the electricity prices and reserve margins included in the past data; and a prediction model construction and inference unit 14 that predicts the electricity market price at the time of prediction using past data for the range including the reserve margin at the time of prediction.

[0066] The electricity market price is generally determined by the balance of supply and demand (supply against demand), but the tendency of this change differs depending on whether the supply and demand is tight or not. Therefore, if the electricity market price is predicted using a single prediction model without considering the tight supply and demand of electricity, the approximation ability will be reduced. In contrast, with the above-mentioned configuration, it is possible to predict the electricity market price using different models for each of multiple intervals of the reserve margin based on the tendency of changes in the electricity market price. For example, it is possible to predict the electricity market price using different models when the supply and demand of electricity is tight (e.g., interval A shown in FIG. 2) and when it is not (e.g., interval B shown in FIG. 2), thereby improving the approximation ability. This makes it possible to respond to fluctuations in the electricity market price when the supply and demand is tight, and to predict the electricity market price with high accuracy.

[0067] The above-described embodiments are intended to facilitate understanding of the present invention and are not intended to limit the present invention. Furthermore, the present invention may be modified or improved without departing from the spirit thereof, and the present invention includes equivalents thereof.

[0068] For example, some of the functions of the information processing device 10 may be executed by another device. For example, in the above-described embodiment, the prediction model construction / inference unit 14 constructs a prediction model, but this is not limiting, and the prediction model construction / inference unit 14 may predict the electricity market price using a prediction model constructed by another device. [Explanation of symbols]

[0069] Information processing device 10 Acquisition part 11 Database 12 Reserve Margin Information Management Department 13 Predictive Model Building and Inference Department 14 Extraction part 141 Correction unit 142 Prediction Department 143 Output section 144

Claims

1. An information processing device for predicting an electricity price at a prediction time, an acquisition unit that acquires historical data including historical electricity prices and related information related to electricity prices, including at least a reserve margin based on electricity demand and electricity supply capacity; an extraction unit that extracts similar data from the past data, the related information of which is similar to the related information at the time of prediction; a correction unit that corrects the electricity price of the similar data based on a difference between a reserve margin included in information related to the similar data and a reserve margin at the time of prediction; a prediction unit that predicts the electricity price at the prediction time based on the corrected electricity price; Including, Information processing device.

2. 2. The information processing device according to claim 1, a setting unit that sets a plurality of intervals of the reserve margin that have different trends in change in the electricity price based on the electricity price and the reserve margin included in the past data, The extraction unit Extract similar data from past data for the interval that includes the reserve margin at the time of forecasting Information processing device.

3. 3. The information processing device according to claim 2, The setting unit Set multiple intervals for the reserve margin using clustering or decision trees Information processing device.

4. 2. The information processing device according to claim 1, The extraction unit extracting a plurality of the similar data; The correction unit correcting the electricity price for each of the plurality of similar data; The prediction unit Predict the electricity price on the forecast date based on the adjusted statistics of multiple electricity prices Information processing device.

5. An information processing device for predicting an electricity price at a prediction time, an acquisition unit that acquires historical data including historical electricity prices and related information related to electricity prices, including at least a reserve margin based on electricity demand and electricity supply capacity; a setting unit that sets a plurality of intervals of the reserve margin that have different trends in change in the electricity price based on the electricity price and the reserve margin included in the past data; a prediction unit that predicts the electricity price at the time of prediction using past data for a section that includes the reserve margin at the time of prediction; Including, Information processing device.

6. An information processing method in which an information processing device predicts an electricity price at a prediction time, obtaining historical data including historical electricity prices and related information related to electricity prices, including at least a reserve margin based on electricity demand and electricity availability; extracting similar data from the past data, the related information of which is similar to the related information at the time of prediction; correcting the electricity price of the similar data based on a difference between the reserve margin included in the related information of the similar data and the reserve margin at the time of prediction; predicting the electricity price at the prediction time based on the corrected electricity price; Including, Information processing methods.

7. An information processing program for predicting an electricity price at a prediction time, On the computer, an acquisition unit that acquires historical data including historical electricity prices and related information related to electricity prices, including at least a reserve margin based on electricity demand and electricity supply capacity; an extraction unit that extracts similar data from the past data, the related information of which is similar to the related information at the time of prediction; a correction unit that corrects the electricity price of the similar data based on a difference between a reserve margin included in information related to the similar data and a reserve margin at the time of prediction; a prediction unit that predicts the electricity price at the prediction time based on the corrected electricity price; To realize Information processing program.

Images