Prediction device, prediction program, prediction system, and method of prediction

Abstract

To readily predict an amount of energy consumption by a novel air conditioner in a short time.SOLUTION: A prediction device (1) includes: an information acquisition unit (151) for acquiring first temperature information and first humidity information; and an index acquisition unit (152) for inputting the first temperature information and the first humidity information into a prediction model (PM) that has mechanically learned with teacher data (TD) to acquire an energy index.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a prediction device, a prediction program, a prediction system, and a prediction method for predicting the amount of energy consumed by air conditioning equipment. [Background technology]

[0002] Technologies for predicting the living environment of a building have been researched and developed. For example, Patent Documents 1 and 2 disclose a living environment simulation system that calculates the comfort level at each position in a building based on a predetermined relationship between the value of an environmental parameter and the comfort level, and the value of the environmental parameter at each position in the building. The value of the environmental parameter is calculated for each position in the building based on the design data of the building and information that affects the indoor environment of the building. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2021-033684 [Patent Document 2] Japanese Patent Application Publication No. 2023-103218 Summary of the Invention [Problem to be solved by the invention]

[0004] However, the technologies disclosed in Patent Documents 1 and 2 calculate the comfort level at the time of acquiring information that affects the indoor environment of a building, but do not calculate the comfort level at any time after the time of acquisition. To begin with, the technologies disclosed in Patent Documents 1 and 2 do not simulate the amount of energy consumed by air conditioning equipment in a building. Furthermore, the technologies disclosed in Patent Documents 1 and 2 require environmental parameters to be calculated in advance before calculating the comfort level, which requires time and effort to output the simulated results.

[0005] An object of one aspect of the present invention is to easily and quickly predict the amount of energy consumed by an air conditioner, which has not been possible in the past. [Means for solving the problem]

[0006] In order to solve the above problem, a prediction device according to one embodiment of the present invention is a prediction device that predicts the amount of energy consumed by air conditioning equipment installed in a target space within real estate, and includes an information acquisition unit that acquires first temperature information indicating the temperature of the target space at a first time point and first humidity information indicating the humidity of the target space at the first time point, and an index acquisition unit that inputs the first temperature information and the first humidity information acquired by the information acquisition unit into a machine-learned prediction model using teacher data that uses first learning temperature information indicating the temperature of the target space at a second time point different from the first time point and first learning humidity information indicating the humidity of the target space at the second time point as example data, and a learning energy index that indicates the amount of energy consumed at a specified time period after the second time point as correct answer data, and acquires an energy index that indicates the prediction result of the amount of energy consumed at a specified time period after the first time point.

[0007] According to the above configuration, the prediction device can perform an unprecedented prediction of the amount of energy consumed by an air conditioner for a predetermined period of time after the first time point, simply by inputting the first temperature information and the first humidity information acquired by the information acquisition unit into the prediction model. This makes it possible to easily and quickly perform an unprecedented prediction of the amount of energy consumed by an air conditioner.

[0008] In one aspect of the prediction device of the present invention, the example data may include a learning visible image in which at least a portion of the target space is captured at the second time point, the information acquisition unit may further acquire a visible image in which at least a portion of the target space is captured at the first time point, and the index acquisition unit may further input the visible image acquired by the information acquisition unit to the prediction model.

[0009] According to the above configuration, the index acquirer can acquire an energy index that takes into account the influence of the state of at least a part of the target space on the thermal sensation of the target space, thereby improving the prediction accuracy of the prediction device.

[0010] In one embodiment of the prediction device of the present invention, the example data includes a learning infrared image of at least a portion of the target space at the second time point, the information acquisition unit further acquires an infrared image of at least a portion of the target space at the first time point, and the index acquisition unit further inputs the infrared image acquired by the information acquisition unit into the prediction model.

[0011] According to this configuration, the index acquisition unit can acquire an energy index that takes into account the influence of temperature distribution in at least a part of the target space on the thermal sensation in the target space, thereby improving the prediction accuracy of the prediction device.

[0012] In one embodiment of the present invention, a prediction device has at least a portion of the outer shape of the target space formed by a target part that is part of the real estate, the example data includes second learning temperature information indicating the temperature of the surface of the target part facing the target space at the second time point, the information acquisition unit further acquires second temperature information indicating the temperature of the surface of the target part facing the target space at the first time point, and the index acquisition unit further inputs the second temperature information acquired by the information acquisition unit into the prediction model.

[0013] According to the above configuration, the index acquirer can acquire an energy index that takes into account the effect of the temperature of the surface of the target part facing the target space on the thermal sensation in the target space, thereby improving the prediction accuracy of the prediction device.

[0014] In one aspect of the present invention, there is provided a prediction device, wherein the real estate includes a plurality of spaces other than the target space, and the example data includes third learning temperature information indicating the temperature at the second time point in at least some of the spaces among the plurality of spaces, fourth learning temperature information indicating the outside air temperature at the second time point, a first learning contribution indicating the degree of influence that the temperature at the second time point in the at least some of the spaces has on the temperature of the target space, and a second contribution indicating the degree of influence that the outside air temperature at the second time point has on the temperature of the target space, and the information acquisition unit acquires the temperature at the first time point in at least some of the spaces among the plurality of spaces. The index acquisition unit may further include a temperature contribution calculation unit that acquires third temperature information indicating the temperature at the first time point and fourth temperature information indicating the outside air temperature at the first time point, and calculates a first contribution indicating the degree of influence that the temperature at the first time point in at least a portion of the space has on the temperature of the target space and a second contribution indicating the degree of influence that the outside air temperature at the first time point has on the temperature of the target space based on the first temperature information, the third temperature information, and the fourth temperature information acquired by the information acquisition unit, and the index acquisition unit may further input each of the first contribution and the second contribution calculated by the temperature contribution calculation unit to the prediction model.

[0015] According to the above configuration, the index acquisition unit can acquire an energy index that takes into account the influence of the temperature distribution of at least some of the spaces among the plurality of spaces on the thermal sensation of the target space and the influence of the outside air temperature on the thermal sensation of the target space, thereby improving the prediction accuracy of the prediction device.

[0016] In a prediction device according to one aspect of the present invention, at least a part of the outer shape of the target space is formed by a target part that is part of the real estate, the ground truth data includes an actual measurement value of an insulation index indicating the insulation performance of the target part at a predetermined time period after the second time point, and the index acquisition unit may further acquire a predicted value of the insulation index of the target part at a predetermined time period after the first time point from the prediction model. With this configuration, the prediction device can predict the insulation index of the target part at a predetermined time period after the first time point, which is a novel prediction of an insulation index not possible in conventional methods.

[0017] The prediction device according to an aspect of the present invention may further include a display control unit that causes a display unit to display at least one of the first contribution and the second contribution calculated by the temperature contribution calculation unit. With this configuration, a user can grasp at least one of the degree of influence that the temperatures of at least some of the spaces among the plurality of spaces have on the temperature of the target space and the degree of influence that the outside air temperature has on the temperature of the target space, simply by visually checking the display unit.

[0018] In one aspect of the present invention, there is provided a prediction device in which a plurality of spaces other than the target space exist within the real estate, and at least a portion of the outline of the target space is formed by a target portion that forms part of the real estate, and the example data includes second learning humidity information indicating the humidity in at least a portion of the plurality of spaces at the second time point, third learning humidity information indicating the humidity outside the real estate at the second time point, a third learning contribution indicating the degree of influence that the humidity in the at least a portion of the spaces at the second time point has on the humidity of the target space, and a fourth learning contribution indicating the degree of influence that the humidity outside the real estate at the second time point has on the humidity of the target space, and the correct answer data includes an actual measurement value of an airtightness index that indicates the airtightness performance of the target portion at a predetermined time after the second time point, and the information acquisition unit is The prediction device may further include a humidity contribution calculation unit that acquires second humidity information indicating the humidity in at least some of the spaces at the first time point and third humidity information indicating the humidity outside the property at the first time point, and calculates a third contribution indicating the degree of influence that the humidity in at least some of the spaces at the first time point has on the humidity of the target space and a fourth contribution indicating the degree of influence that the humidity outside the property at the first time point has on the humidity of the target space based on the first humidity information, the second humidity information, and the third humidity information acquired by the information acquisition unit, wherein the index acquisition unit further inputs the third contribution and the fourth contribution calculated by the humidity contribution calculation unit into the prediction model to further acquire a predicted value of the airtightness index for the target location at a predetermined time period after the first time point. According to the above configuration, the prediction device can predict the value of the airtightness index for the target location at a predetermined time period after the first time point, which is a novel method of predicting the value of the airtightness index.

[0019] The prediction device according to each aspect of the present invention may be realized by a computer. In this case, a control program for the prediction device that causes a computer to operate as each unit (software element) of the prediction device to realize the prediction device, and a computer-readable recording medium on which the control program is recorded, also fall within the scope of the present invention.

[0020] In order to solve the above-mentioned problem, a prediction system according to one embodiment of the present invention is a prediction system that predicts the amount of energy consumed by air conditioning equipment installed in a target space within real estate, and includes an information acquisition unit that acquires first temperature information indicating the temperature of the target space at a first time point and first humidity information indicating the humidity of the target space at the first time point, and an index acquisition unit that inputs the first temperature information and the first humidity information acquired by the information acquisition unit into a machine-learned prediction model using teacher data that uses first learning temperature information indicating the temperature of the target space at a second time point different from the first time point and first learning humidity information indicating the humidity of the target space at the second time point as example data, and a learning energy index that indicates the amount of energy consumed at a specified time period after the second time point as correct answer data, and acquires an energy index that indicates the prediction result of the amount of energy consumed at a specified time period after the first time point.

[0021] In order to solve the above-mentioned problem, a prediction method according to one embodiment of the present invention is a prediction method executed by one or more computers to predict the amount of energy consumed by air conditioning equipment installed in a target space within real estate, and includes an information acquisition step of acquiring first temperature information indicating the temperature of the target space at a first time point and first humidity information indicating the humidity of the target space at the first time point, and an index acquisition step of inputting the first temperature information and the first humidity information acquired in the information acquisition step into a machine-learned prediction model using teacher data in which first learning temperature information indicating the temperature of the target space at a second time point different from the first time point and first learning humidity information indicating the humidity of the target space at the second time point are used as example data, and a learning energy index indicating the amount of energy consumed at a specified time period after the second time point is used as correct answer data, to acquire an energy index indicating the predicted result of the amount of energy consumed at a specified time period after the first time point. [Effects of the Invention]

[0022] According to one aspect of the present invention, it is possible to easily and quickly predict the amount of energy consumed by an air conditioner, which has not been possible in the past. [Brief explanation of the drawings]

[0023] [Figure 1] 1 is a block diagram showing a functional configuration of a prediction system according to a first embodiment of the present invention. [Figure 2] 1 is a diagram showing the floor plan of a building to which a prediction system according to one embodiment of the present invention is applied. [Figure 3] 1 is a flowchart showing a flow of processing executed by the prediction system according to the first embodiment of the present invention. [Figure 4] FIG. 10 is a block diagram showing the functional configuration of a prediction system according to a second embodiment of the present invention. [Figure 5] 5 is a graph showing the first contribution degree of room temperature and the second contribution degree of outdoor air temperature for each room, displayed on a display unit of the prediction device shown in FIG. 4. [Figure 6]10 is a flowchart showing the flow of processing executed by a prediction system according to a second embodiment of the present invention. [Figure 7] FIG. 10 is a block diagram showing the functional configuration of a prediction system according to a third embodiment of the present invention. [Figure 8] 10 is a flowchart showing the flow of processing executed by a prediction system according to a third embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0024] First to third embodiments of the present invention will be described in detail below with reference to FIGS. 1 to 8. In the first to third embodiments, a building, specifically a detached house, will be described as an example of real estate to which a prediction system according to one aspect of the present invention is applied. Note that real estate to which a prediction system according to one aspect of the present invention is applied is not limited to buildings, and may be any real estate having one or more internal spaces. In addition, registered ships, registered automobiles, and automobiles with mortgages are also applicable to the prediction system according to one aspect of the present invention. However, vacant land is not applicable because it is not possible to conceive of an internal space.

[0025] [First embodiment] <Configuration of the prediction system> The configuration of a prediction system 4 according to a first embodiment of the present invention will be described with reference to FIGS. 1 and 2. The prediction system 4 is a system that predicts the amount of energy consumed by air conditioning equipment (not shown) installed in a target space SS in a building B. The target space SS is a space in the building B in which the air conditioning equipment is installed. In this embodiment, the room "Room 1" shown in FIG. 2 is taken as the target space SS. The air conditioning equipment is, for example, an air conditioner, a humidifier, a dehumidifier, a ventilation fan, etc., and the prediction system 4 can predict the amount of energy consumed by any type of air conditioning equipment.

[0026] As shown in Fig. 1, the prediction system 4 includes a prediction device 1, a temperature and humidity sensor 2, and a server 3. The prediction device 1, the temperature and humidity sensor 2, and the server 3 are communicatively connected via a network NW. The network NW may include, for example, a wired local area network (LAN), a wireless LAN, a wide area network (WAN), or the like, but is not limited to these. Furthermore, while the example of Fig. 1 shows one prediction device 1, one temperature and humidity sensor 2, and one server 3, the number of each device included in the prediction system 4 is not limited to one.

[0027] (Prediction device) The prediction device 1 is an information processing device that predicts the energy consumption amount of the air conditioning equipment described above. In the first to third embodiments, the prediction device 1 is assumed to be a user terminal such as a stationary personal computer, a tablet terminal, or a smartphone. The prediction device 1 includes a display unit 11, an input unit 12, a storage unit 13, a communication unit 14, and a control unit 15.

[0028] The display unit 11 is an output unit that displays various information. The input unit 12 accepts various operations from the user. The display unit 11 and the input unit 12 may be, but are not limited to, a touch panel formed integrally. The display unit 11 may be configured, for example, by a display. The input unit 12 may be configured, for example, by a keyboard, a touchpad, or a mouse. The storage unit 13 is configured, for example, by a memory, and stores various information referenced by the control unit 15. The display unit 11, the input unit 12, and the storage unit 13 may each be connected to the prediction device 1 as peripheral devices.

[0029] The communication unit 14 communicates with other devices via the network NW. The communication unit 14 is configured, for example, by a communication interface. The control unit 15 controls all the units of the prediction device 1, and is realized, for example, by a processor executing a terminal program. The terminal program constitutes at least a part of a program according to one aspect of the present invention. The control unit 15 includes an information acquisition unit 151 and an index acquisition unit 152.

[0030] The information acquisition unit 151 acquires first temperature information indicating the temperature of the target space SS at a first time point and first humidity information indicating the humidity of the target space SS at the first time point. In the first to third embodiments, the first time point is the time point when the temperature and humidity sensor 2 starts measurement, but this is not limited to this. For example, the first time point may be the time point when a certain amount of time (seconds, minutes, hours) has passed or a certain number of days has passed since the temperature and humidity sensor 2 started measurement.

[0031] The index acquiring unit 152 acquires energy indices by inputting the first temperature information and the first humidity information acquired by the information acquiring unit 151 into the prediction model PM. The prediction model PM is a learning model that has undergone machine learning using the training data TD. Examples of machine learning (supervised learning) for generating the prediction model PM include various well-known techniques such as decision trees, regression analysis, support vector machines, random forests, and neural networks.

[0032] The training data TD uses, as example data, first learning temperature information indicating the temperature of the target space SS at a second time point different from the first time point, and first learning humidity information indicating the humidity of the target space SS at the second time point. The training data TD also uses, as correct answer data, a learning energy index indicating the amount of energy consumed by the air conditioning equipment at a predetermined time period after the second time point.

[0033] The second time point may be any time point different from the first time point. For example, the second time point may be several hours, several days, or several weeks before the time point (first time point) when the temperature and humidity sensor 2 starts measurement. Furthermore, for example, the training data TD may be composed of a plurality of example data and a plurality of correct answer data, with the second time point being every few seconds, every few minutes, every few hours, or every day during a certain period before the time point (first time point) when the temperature and humidity sensor 2 starts measurement.

[0034] The energy index is output data from the prediction model PM and is an index that indicates the prediction result of the amount of energy consumed by the air conditioning equipment at a predetermined time period after the first time point. In other words, when the first temperature information and the first humidity information acquired by the information acquisition unit 151 are used as explanatory variables of the prediction model PM, the energy index is the objective variable of the prediction model PM.

[0035] The predetermined period after the first time point can be any period after the first time point. For example, the predetermined period after the first time point may be several seconds, several minutes, several hours, or several days after the temperature and humidity sensor 2 starts measurement (first time point). For example, the predetermined period after the first time point may be several tens of minutes after 10 minutes, several hours after one hour, or several days after one day after the temperature and humidity sensor 2 starts measurement (first time point). For example, if the training data TD is composed of multiple example data sets and multiple correct answer data sets, each set at a time point every few seconds, which is set as the second time point, during a certain period of time after the temperature and humidity sensor 2 starts measurement (first time point), the predetermined period after the first time point may be several days after the first time point or several hours after the first time point.

[0036] The amount of energy consumption that is the target of the energy index may be the amount of power consumed by the air conditioner during a predetermined period of time after the first point in time. Alternatively, the amount of energy consumption may be the amount of electrical energy consumed per second by the air conditioner during a predetermined period of time after the first point in time, i.e., the power consumption.

[0037] There are no particular limitations on the expression format of the energy index. For example, an energy index for an amount of energy consumption within a normal range may be expressed as "normal," an energy index for an amount less than normal may be expressed as "saving," and an energy index for an amount more than normal may be expressed as "waste." In this case, a reference value for the amount of energy consumption used to determine whether the energy index will be "normal" or "saving," and a reference value for the amount of energy consumption used to determine whether the energy index will be "normal" or "waste," are set in advance. Then, by comparing the predicted value of the amount of energy consumption predicted by the prediction model PM with these reference values, the prediction model PM may output an energy index of "normal," "saving," or "waste."

[0038] Each of the above-mentioned reference values ​​may be appropriately changed to a value according to the season predicted by the prediction system 4. Furthermore, "general" may be "△" or "2", "saving" may be "○" or "1", and "progress" may be "×" or "3". Furthermore, the energy index may be expressed in multiple stages, such as four or more stages, instead of three stages as in the above example. Alternatively, the energy index may be a predicted value of the amount of energy consumption.

[0039] The learning energy index is an index that indicates the amount of energy consumed (actual measured value) by the air conditioner at a predetermined time period after the second time point. In other words, the learning energy index is the objective variable of the prediction model PM when the first learning temperature information and the first learning humidity information are used as explanatory variables of the prediction model PM.

[0040] Regarding the predetermined period after the second time point, various periods after the second time point are conceivable. For example, the predetermined period after the second time point may be a period several seconds, several minutes, several hours, or several days after the second time point. Furthermore, for example, the predetermined period after the second time point may be several tens of minutes after 10 minutes have passed, several hours after one hour has passed, or several days after one day has passed. The amount of consumed energy that is the subject of the learning energy index and the expression format of the learning energy index are the same as those of the energy index.

[0041] (Temperature and humidity sensor) The temperature and humidity sensor 2 is a sensor that functions both as a temperature sensor that measures temperature and as a humidity sensor that measures humidity, and measures the temperature and humidity of the target space SS. In the first to third embodiments, the temperature and humidity sensor 2 measures the temperature and humidity of the target space SS at a first point in time, and the temperature and humidity of the target space SS at a second point in time. There are no particular limitations on where the temperature and humidity sensor 2 is placed, but it is preferable to place the temperature and humidity sensor 2 in a location that will increase measurement accuracy depending on the shape and size of the target space SS, the placement position of the target space SS in building B, the overall structure of building B, etc.

[0042] It is not essential that the prediction system 4 includes the temperature and humidity sensor 2, and any measuring device capable of measuring the temperature and humidity of the target space SS may be included in the prediction system 4. Alternatively, instead of the temperature and humidity sensor 2, two sensors, a temperature sensor and a humidity sensor, may be included in the prediction system 4. Furthermore, although one temperature and humidity sensor 2 is shown in the example of FIG. 1, multiple temperature and humidity sensors 2 may be included in the prediction system 4.

[0043] (server) The server 3 is an information processing device configured to be able to communicate with each of the prediction device 1 and the temperature and humidity sensor 2 via a network NW. The server 3 includes a storage unit 31, a communication unit 32, and a control unit 33. The storage unit 31 is configured, for example, by a memory, and stores various information referenced by the control unit 33. The storage unit 31 may be connected to the server 3 as a peripheral device.

[0044] The storage unit 31 also stores the teacher data TD and the prediction model PM. In the first to third embodiments, both the teacher data TD and the prediction model PM are generated in advance and stored in the storage unit 31, but this is not limiting. For example, the server 3 may include at least one of a teacher data generation unit that generates the teacher data TD and a model generation unit that generates the prediction model PM. In this case, the server 3 acquires the first learning temperature information and the first learning humidity information from the temperature and humidity sensor 2. The teacher data generation unit then stores the generated teacher data TD in the storage unit 31. Alternatively, the model generation unit stores the generated prediction model PM in the storage unit 31. The prediction device 1 may include at least one of a teacher data generation unit and a model generation unit.

[0045] Furthermore, it is not essential that the storage unit 31 stores the training data TD and the prediction model PM. For example, the storage unit 31 may store only one of the training data TD and the prediction model PM. In this case, the storage unit 13 may store the other, or another device not included in the prediction system 4 may store the other.

[0046] The communication unit 32 communicates with other devices via the network NW. The communication unit 32 is configured, for example, by a communication interface. The control unit 15 controls all the units of the server 3, and is realized, for example, by a processor executing a server program. The server program constitutes at least a part of a program according to one aspect of the present invention.

[0047] <Process flow executed by the prediction system> The flow of processing executed by the prediction system 4 will be described with reference to Fig. 3. Fig. 3 is a flowchart showing the flow of processing executed by the prediction system 4. This flowchart includes each step of the prediction method according to the first embodiment of the present invention.

[0048] In S11, the temperature and humidity sensor 2 measures the temperature of the target space SS at a first time point and transmits it as first temperature information to the prediction device 1. The temperature and humidity sensor 2 also measures the humidity of the target space SS at the first time point and transmits it as first humidity information to the prediction device 1. The transmission timing of the first temperature information and the first humidity information may be synchronized or asynchronous.

[0049] In S12 (information acquisition step), the information acquisition unit 151 acquires first temperature information and first humidity information. Specifically, the communication unit 14 transmits the first temperature information and first humidity information received from the temperature and humidity sensor 2 to the information acquisition unit 151, and the information acquisition unit 151 acquires the first temperature information and first humidity information. Then, the information acquisition unit 151 transmits the acquired first temperature information and first humidity information to the index acquisition unit 152. The transmission timing of the first temperature information and the first humidity information may be synchronous or asynchronous.

[0050] In S13, the index acquisition unit 152 transmits an output command to the server 3. The output command is a signal that causes the server 3 to transmit the prediction model PM stored in the storage unit 31 to the index acquisition unit 152. Note that the order of processing in S13 is not limited to the example in FIG. 3. For example, the processing in S13 may be executed before the processing in S11, or the processing in S13 may be executed between the processing in S11 and the processing in S12.

[0051] In S14, the server 3 transmits the prediction model PM to the prediction device 1. Specifically, the communication unit 32 transmits the output command received from the index acquisition unit 152 to the control unit 33, causing the control unit 33 to read out the prediction model PM from the storage unit 31. Then, the control unit 33 transmits the read-out prediction model PM to the prediction device 1 via the communication unit 32.

[0052] In S15 (index acquisition step), the index acquisition unit 152 acquires energy indices. Specifically, the index acquisition unit 152 inputs the first temperature information and the first humidity information received from the information acquisition unit 151 into the prediction model PM received from the server 3 via the communication unit 14. Then, the index acquisition unit 152 acquires the energy indices output from the prediction model PM. When the processing of S15 ends, the entire series of processes executed by the prediction system 4 ends. Note that the index acquisition unit 152 may cause the display unit 11 to display the energy indices acquired from the prediction model PM.

[0053] <Modification> (Example data and input data variations) In order to improve the prediction accuracy of the prediction model PM, various information other than the first learning temperature information and the first learning humidity information may be included in the example data of the training data TD. For the same purpose, various information other than the first temperature information and the first humidity information may be used as input data (explanatory variables) to the prediction model PM.

[0054] For example, the example data of the training data TD may further include a training visible image. The training visible image is a visible image of at least a portion of the target space SS at a second time point. In this case, the information acquisition unit 151 further acquires a visible image of at least a portion of the target space SS at a first time point (hereinafter, an input visible image) and transmits it to the index acquisition unit 152. The index acquisition unit 152 then inputs the first temperature information, the first humidity information, and the input visible image received from the information acquisition unit 151 into the prediction model PM to acquire an energy index with improved prediction accuracy. Note that the training visible image and the input visible image are preferably visible images of the entire target space SS captured by a 360-degree camera or the like.

[0055] Furthermore, for example, the example data of the training data TD may further include a training infrared image. The training infrared image is an infrared image of at least a portion of the target space SS captured at a second time point. In this case, the information acquisition unit 151 further acquires an infrared image of at least a portion of the target space SS captured at a first time point (hereinafter, an input infrared image) and transmits it to the index acquisition unit 152. The index acquisition unit 152 then inputs the first temperature information, the first humidity information, and the input infrared image received from the information acquisition unit 151 into the prediction model PM to acquire an energy index with improved prediction accuracy. Note that the training infrared image and the input infrared image are preferably infrared images of the entire target space SS captured using a 360° camera or the like.

[0056] When using learning infrared images and input infrared images, the solar radiation of the target location may also be used as example data for the training data TD and as input data for the prediction model PM. Using infrared images and solar radiation together further improves the prediction accuracy of energy indices. Note that the type of target location for measuring solar radiation is not particularly limited, but it is preferable to select a type of measurement target that improves the measurement accuracy of solar radiation. Furthermore, it is not necessary to use solar radiation together with infrared images; for example, solar radiation may be used instead of infrared images. In this case, the prediction accuracy of energy indices will also improve.

[0057] Also, for example, the example data of the training data TD may further include second learning temperature information. The second learning temperature information is information indicating the temperature of the surface of the target part facing the target space SS at a second time point. The target part is a part that forms part of the building B and forms at least part of the outline of the target space SS. Examples of the target part include a ceiling panel, a wall, a floor, etc.

[0058] In this case, the information acquiring unit 151 further acquires second temperature information indicating the temperature of the surface of the target part on the target space SS side at the first time point, and transmits the second temperature information to the index acquiring unit 152. Then, the index acquiring unit 152 inputs each of the first temperature information, the second temperature information, and the first humidity information transmitted from the information acquiring unit 151 into the prediction model PM, and acquires an energy index with improved prediction accuracy.

[0059] The example data of the training data TD may additionally include two or more of a training visible image, a training infrared image, or second training temperature information. In this case, in addition to the first temperature information and first humidity information, two or more of an input visible image, an input infrared image, or second temperature information are input to the prediction model PM, further improving the prediction accuracy of the energy index.

[0060] (Variations of the correct answer data and output data) The index acquiring unit 152 may further acquire various indexes other than the energy index from the prediction model PM. That is, the prediction model PM may further output various indexes other than the energy index as output data (objective variables).

[0061] For example, the prediction model PM may further output, as output data, a discomfort index indicating the predicted result of the discomfort index of the target space SS at a predetermined time period after the first time point. In this case, the correct answer data of the training data TD will further include a learning discomfort index indicating the discomfort index (actual measured value) of the target space SS at a predetermined time period after the second time point. Meanwhile, the input data (explanatory variables) to the prediction model PM may be the first temperature information and the first humidity information, as in the first embodiment.

[0062] For example, the prediction model PM may further output, as output data, a thickness index indicating the predicted result of the thickness index (wet bulb globe temperature) of the target space SS at a predetermined time after the first time point. In this case, the example data of the training data TD will further include learning radiation temperature information indicating the temperature of radiant heat in the target space SS at the second time point. The correct answer data of the training data TD will further include a learning thickness index indicating the thickness index (actual measured value) of the target space SS at a predetermined time after the second time point. Furthermore, the input data (explanatory variables) to the prediction model PM will further include radiation temperature information indicating the temperature of radiant heat in the target space SS at the first time point.

[0063] Second Embodiment A second embodiment of the present invention will be described below. For ease of explanation, the same reference numerals will be used to designate components having the same functions as those described in the first embodiment, and their descriptions will not be repeated. This also applies to a third embodiment described below.

[0064] <Configuration of the prediction system> The configuration of a prediction system 4A according to a second embodiment of the present invention will be described with reference to Figures 2, 4, and 5. As shown in Figure 4, the prediction system 4A differs from the prediction system 4 in that it includes a prediction device 1A instead of the prediction device 1. Here, Figure 4 is a block diagram showing the functional configuration of the prediction system 4A, but since the temperature and humidity sensor 2 and server 3 are the same as those in the prediction system 4, they are not shown in the figure to simplify the drawing. The same applies to the block diagram in Figure 7.

[0065] Prediction device 1A differs from prediction device 1 in that it includes control unit 15A instead of control unit 15. Control unit 15A includes information acquisition unit 151A, temperature contribution calculation unit 153, index acquisition unit 152A, and display control unit 154.

[0066] In addition to the first temperature information and the first humidity information, the information acquisition unit 151A further acquires third temperature information and fourth temperature information. The third temperature information is information indicating the temperature at a first point in time in at least some of the spaces other than the target space SS that exist in building B. The fourth temperature information is information indicating the outside air temperature at the first point in time. Hereinafter, at least some of the spaces other than the target space SS that exist in building B will be referred to as "non-target spaces."

[0067] In this embodiment, all spaces within building B are defined as rooms "Room 1" to "Room 5," the kitchen "Kitchen," the earthen floor "Doma," and the veranda "Engawa" shown in FIG. 2, as well as the foundation of building B ("Foundation" in FIG. 5) not shown in FIG. 2. A space to be a target space SS is arbitrarily selected from all spaces within building B, and the temperature and humidity of the selected space at a first time point are measured by the temperature and humidity sensor 2. The temperature measured by this temperature and humidity sensor 2 becomes the first temperature information, and the humidity becomes the first humidity information. Furthermore, a non-target space is arbitrarily selected from all spaces within building B, and the temperature of the selected space at a first time point is measured by another temperature and humidity sensor 2. The temperature measured by this temperature and humidity sensor 2 becomes the third temperature information. Furthermore, the temperature at a point "T_out" outside building B shown in FIG. 2 at a first time point is measured by another temperature and humidity sensor 2. The temperature measured by this temperature and humidity sensor 2 becomes the fourth temperature information.

[0068] Regarding the selection of the target space SS and the non-target space described above, for example, the selection contents of the target space SS and the non-target space may be stored in advance in the storage unit 13 or 31. Also, for example, the input unit 12 may accept an input operation of each selection content, and the control unit 15A may set each of the target space SS and the non-target space.

[0069] The temperature contribution calculation unit 153 calculates the first contribution and the second contribution based on the first temperature information, the third temperature information, and the fourth temperature information acquired by the information acquisition unit 151A. The first contribution is an index indicating the degree of influence that the temperature in the non-target space at the first time point has on the temperature of the target space SS. The second contribution is an index indicating the degree of influence that the outside air temperature at the point "T_out" at the first time point has on the temperature of the target space SS.

[0070] In this embodiment, the temperature contribution calculation unit 153 sets the Feature Importance (an index for evaluating the classification contribution rate of a target in a feature amount) calculated using Gini impurity as the first and second contributions. This also applies to the case where the humidity contribution calculation unit 155, which will be described later, calculates the third and fourth contributions.

[0071] The index acquiring unit 152A inputs the first temperature information and the first humidity information, as well as the first contribution degree and the second contribution degree calculated by the temperature contribution degree calculating unit 153, to the prediction model PM according to this embodiment. Then, the index acquiring unit 152A acquires an energy index with improved prediction accuracy from the prediction model PM according to this embodiment.

[0072] Here, the prediction model PM according to this embodiment is generated by machine learning using the training data TD according to this embodiment. The example data of the training data TD according to this embodiment includes, in addition to the first learning temperature information and the first learning humidity information, third learning temperature information, fourth learning temperature information, a first learning contribution, and a second learning contribution. The third learning temperature information is information indicating the temperature in the non-target space at a second time point. The fourth learning temperature information is information indicating the outside air temperature at the point "T_out" at the second time point. The first learning contribution is an index indicating the degree of influence that the temperature in the non-target space at the second time point has on the temperature of the target space SS. The second learning contribution is an index indicating the degree of influence that the outside air temperature at the point "T_out" at the second time point has on the temperature of the target space SS.

[0073] In this embodiment, the methods for acquiring the first learning temperature information, the first learning humidity information, the third learning temperature information, and the fourth learning temperature information are the same as the methods for acquiring the first temperature information, the first humidity information, the third temperature information, and the fourth temperature information described above. The first learning contribution and the second learning contribution temperature are calculated by the temperature contribution calculation unit 153 based on the first learning temperature information, the third learning temperature information, and the fourth learning temperature information acquired by the information acquisition unit 151A.

[0074] The display control unit 154 displays at least one of the first contribution and the second contribution calculated by the temperature contribution calculation unit 153 on the display unit 11. Display control by the display control unit 154 will be described below using the example of FIG. 5. In the example of FIG. 5, the first contribution and the second contribution are graphed and displayed collectively on the display unit 11 when the target space SS is (i) room "Room 1," (ii) room "Room 2," (iii) room "Room 3," and (iv) room "Room 4." Furthermore, to calculate the first contribution and the second contribution, four types of prediction models PM were used, which were obtained using two types of machine learning methods (learning methods 1 and 2) and two types of training data TD (training data 1 and 2). Here, learning method 1 is a case where an optimization algorithm is used, and learning method 2 is a case where ensemble learning is used.

[0075] In the case of (i), the rooms "Room2" and "Room3," the dirt floor "Doma," the foundation "Foundation," and the location "T_out" are selected as excluded spaces (the graph for "Room1" in Figure 5). In the case of (ii), the room "Room1," the dirt floor "Doma," the foundation "Foundation," and the location "T_out" are selected as excluded spaces (the graph for "Room2" in Figure 5). In the case of (iii), the rooms "Room1" and "Room4," the veranda "Engawa," the foundation "Foundation," and the location "T_out" are selected as excluded spaces (the graph for "Room3" in Figure 5). In the case of (iv), the room "Room3," the veranda "Engawa," the foundation "Foundation," and the location "T_out" are selected as excluded spaces (the graph for "Room4" in Figure 5).

[0076] As shown in the example of Figure 5, by displaying and visualizing the calculation results (at least one of the first contribution and the second contribution) of the temperature contribution calculation unit 153 on the display unit 11, it is possible to easily understand the extent to which the temperature and humidity of the non-target space affect the predicted results of the amount of energy consumption.

[0077] The display manner of the first contribution degree and the second contribution degree on the display unit 11 is not limited to the example in Fig. 5. Furthermore, if the server 3 is provided with a display unit, the display control unit 154 may cause the display unit to display at least one of the first contribution degree and the second contribution degree. Alternatively, the display control unit 154 may cause the display unit of another device that is not part of the prediction system 4A to display at least one of the first contribution degree and the second contribution degree. Furthermore, the prediction device 1A may not be provided with the display control unit 154.

[0078] <Process flow executed by the prediction system> The flow of processing executed by the prediction system 4A will be described with reference to Figure 6. In S21, the temperature and humidity sensor 2 measures the temperature of the non-target space at a first time point and transmits it to the prediction device 1A as third temperature information. In addition, another temperature and humidity sensor 2 measures the outside air temperature at the location "T_out" at the first time point and transmits it to the prediction device 1A as fourth temperature information. The measurement and transmission of the first temperature information and first humidity information are the same as in S11. The transmission timing of the first temperature information, first humidity information, third temperature information, and fourth temperature information may be partially or completely synchronized or asynchronous.

[0079] In S22, the information acquisition unit 151A acquires first temperature information, first humidity information, third temperature information, and fourth temperature information. Details of the method for acquiring this information are the same as in S12. Then, the information acquisition unit 151A transmits the acquired first temperature information and first humidity information to the index acquisition unit 152A. Furthermore, the information acquisition unit 151A transmits the acquired first temperature information, third temperature information, and fourth temperature information to the temperature contribution calculation unit 153. The timing of transmitting the first temperature information, first humidity information, third temperature information, and fourth temperature information may be partially or entirely synchronized or asynchronous.

[0080] In S23, the temperature contribution calculation unit 153 calculates the first contribution and the second contribution. Then, the temperature contribution calculation unit 153 transmits the calculated first contribution and the second contribution to the index acquisition unit 152A and the display control unit 154. The process of S24 is the same as that of S13. The process of S25 is the same as that of S14.

[0081] In S26, the index acquisition unit 152A inputs the first temperature information and the first humidity information received from the information acquisition unit 151A and the first contribution and the second contribution received from the temperature contribution calculation unit 153 into the prediction model PM according to this embodiment to acquire an energy index. In S27, the display control unit 154 causes the display unit 11 to display at least one of the first contribution and the second contribution. Specifically, the display control unit 154 transmits the first contribution and the second contribution acquired from the temperature contribution calculation unit 153 to the display unit 11, thereby controlling the display of the display unit 11 so that the display unit 11 displays the first contribution and the second contribution. Note that the order of processing in S27 is not limited to the example of FIG. 6. The processing in S27 may be executed at any timing between after the processing in S23 and before the processing in S26. When the processing in S27 ends, the entire series of processing executed by the prediction system 4A ends.

[0082] <Modification> The index acquisition unit 152A may further acquire various indexes other than the energy index from the prediction model PM according to this embodiment. That is, the prediction model PM according to this embodiment may further output various indexes other than the energy index as output data (objective variables).

[0083] For example, the prediction model PM according to this embodiment may further output, as output data, a predicted value of an insulation index indicating the insulation performance of the target location at a predetermined time period after the first time point. Examples of the insulation index include a normal heat transfer coefficient or a heat loss coefficient (Q value). In this case, the correct answer data of the training data TD according to this embodiment will further include an actual measurement value of the target location at a predetermined time period after the second time point, such as the normal heat transfer coefficient, which is an insulation index. Meanwhile, the input data (explanatory variables) to the prediction model PM may be the first temperature information, the first humidity information, the first contribution rate, and the second contribution rate, as in the second embodiment.

[0084] Third Embodiment <Configuration of the prediction system> The configuration of a prediction system 4B according to a third embodiment of the present invention will be described with reference to Fig. 7 and Fig. 8. As shown in Fig. 7, the prediction system 4B differs from the prediction systems 4 and 4A in that it includes a prediction device 1B instead of the prediction devices 1 and 1A. The prediction device 1B differs from the prediction devices 1 and 1A in that it includes a control device 15B instead of the control devices 15 and 15A. The control device 15B includes an information acquisition unit 151B, a humidity contribution calculation unit 155, and an index acquisition unit 152B.

[0085] In addition to the first temperature information and the first humidity information, the information acquisition unit 151B further acquires second humidity information and third humidity information. The second humidity information is information indicating the humidity in the non-target space at a first time point. The third humidity information is information indicating the humidity outside the building B at the first time point. That is, in this embodiment, the temperature and humidity sensors 2 are installed in both the non-target space and the outside of the building B.

[0086] The humidity contribution calculation unit 155 calculates the third contribution and the fourth contribution based on the first humidity information, the second humidity information, and the third humidity information acquired by the information acquisition unit 151B. The third contribution is an index indicating the degree of influence that the humidity in the non-target space at the first time point has on the humidity in the target space SS. The fourth contribution is an index indicating the degree of influence that the humidity outside the building B at the first time point has on the humidity in the target space SS.

[0087] The index acquiring unit 152B inputs the third and fourth contributions calculated by the humidity contribution calculating unit 155, in addition to the first temperature information and the first humidity information, into the prediction model PM according to this embodiment. Then, the index acquiring unit 152B acquires an energy index with improved prediction accuracy from the prediction model PM according to this embodiment. The index acquiring unit 152B also acquires a predicted value of an airtightness index indicating the airtightness performance of the target portion at a predetermined time period after the first time point in the target portion, from the prediction model PM according to this embodiment. The airtightness index may be, for example, the gap area, or may be a gap equivalent area (C value) obtained by dividing the gap area of ​​the entire target portion by the total floor area.

[0088] Here, the prediction model PM according to this embodiment is generated by machine learning using the training data TD according to this embodiment. The example data of the training data TD according to this embodiment includes, in addition to the first training temperature information and the first training humidity information, second training humidity information, third training humidity information, third training contribution, and fourth training contribution. The second training humidity information is information indicating the humidity in the non-target space at a second time point. The third training humidity information is information indicating the humidity outside of building B at a second time point. The third training contribution is an index indicating the degree of influence that the humidity in the non-target space at a second time point has on the temperature of the target space SS. The fourth training contribution is an index indicating the degree of influence that the humidity outside of building B at a second time point has on the temperature of the target space SS.

[0089] In this embodiment, the methods for acquiring the second and third learning humidity information are the same as those for acquiring the second and third humidity information described above. The third and fourth learning contribution temperatures are calculated by the humidity contribution calculation unit 155 based on the first, second, and third learning humidity information acquired by the information acquisition unit 151B.

[0090] <Process flow executed by the prediction system> The flow of processing executed by prediction system 4B will be described with reference to Figure 8. In S31, the temperature and humidity sensor 2 measures the humidity in a non-target space at a first time point and transmits the measured humidity to prediction device 1B as second humidity information. In addition, another temperature and humidity sensor 2 measures the humidity outside building B at the first time point and transmits the measured humidity to prediction device 1B as third humidity information. The measurement and transmission of the first temperature information and first humidity information are the same as in S11. The timing of transmission of the first temperature information, first humidity information, second humidity information, and third humidity information may be partially or entirely synchronized or asynchronous.

[0091] In S32, the information acquisition unit 151B acquires first temperature information, first humidity information, second humidity information, and third humidity information. Details of the method for acquiring this information are the same as in S12. Then, the information acquisition unit 151B transmits the acquired first temperature information and first humidity information to the index acquisition unit 152B. Furthermore, the information acquisition unit 151B transmits the acquired first humidity information, second humidity information, and third humidity information to the humidity contribution calculation unit 155. The timing of transmitting the first temperature information, first humidity information, second humidity information, and third humidity information may be partially or entirely synchronized or asynchronous.

[0092] In S33, the humidity contribution calculation unit 155 calculates the third contribution and the fourth contribution. Then, the humidity contribution calculation unit 155 transmits the calculated third contribution and the fourth contribution to the index acquisition unit 152B. The process of S34 is the same as S13. The process of S35 is the same as S14.

[0093] In S36, the index acquisition unit 152B inputs the first temperature information and first humidity information received from the information acquisition unit 151B and the third contribution and fourth contribution received from the humidity contribution calculation unit 155 into the prediction model PM according to this embodiment to acquire an energy index. The index acquisition unit 152B also inputs the aforementioned information into the prediction model PM according to this embodiment to acquire a predicted value of the airtightness index for the target location at a predetermined time period after the first time point. The end of the process of S36 marks the end of the entire series of processes executed by the prediction system 4B. The index acquisition unit 152B may display at least one of the energy index and the aforementioned predicted value acquired from the prediction model PM according to this embodiment on the display unit 11.

[0094] [Software implementation example] The functions of the prediction device 1 (hereinafter referred to as the "device") can be realized by a program for causing one or more computers to function as the device, and a program for causing one or more computers to function as each control block of the device (particularly each part included in the control unit 15).

[0095] In this case, the device includes a computer having at least one control device (e.g., a processor) and at least one storage device (e.g., a memory) as hardware for executing the program. The control device and storage device execute the program to realize each function described in each of the above embodiments.

[0096] The program may be non-transitory and may be recorded on one or more computer-readable recording media. The recording media may or may not be included in the device. In the latter case, the program may be supplied to the device via any wired or wireless transmission medium.

[0097] In addition, some or all of the functions of each control block can be realized by a logic circuit. For example, an integrated circuit in which a logic circuit that functions as each control block is formed is also included in the scope of the present invention. In addition, the functions of each control block can be realized by, for example, a quantum computer.

[0098] [Additional Notes] The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. Embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present invention. [Explanation of symbols]

[0099] 1, 1A, 1B Predictor 4, 4A, 4B Prediction Systems 151, 151A, 151B Information acquisition section 152, 152A, 152B Index acquisition part 153 Temperature contribution calculation unit 154 Display control unit 155 Humidity contribution calculation unit B. Building (real estate) PM Prediction Model SS Target Space TD teacher data

Claims

1. A prediction device that predicts the amount of energy consumed by air conditioning equipment installed in a target space within real estate, an information acquisition unit that acquires first temperature information indicating a temperature of the target space at a first time point and first humidity information indicating a humidity of the target space at the first time point; a prediction device comprising: an index acquisition unit that inputs the first temperature information and the first humidity information acquired by the information acquisition unit into a machine-learned prediction model using teacher data in which first learning temperature information indicating the temperature of the target space at a second time point different from the first time point and first learning humidity information indicating the humidity of the target space at the second time point are used as example data, and a learning energy index indicating the amount of energy consumed at a specified time point after the second time point is used as correct answer data, and acquires an energy index indicating the prediction result of the amount of energy consumed at a specified time point after the first time point.

2. the example data includes a learning visible image obtained by capturing at least a part of the target space at the second time point; The information acquisition unit further acquires a visible image in which at least a portion of the target space is captured at the first time point, The prediction device according to claim 1 , wherein the index acquisition unit further inputs the visible image acquired by the information acquisition unit into the prediction model.

3. the example data includes a learning infrared image in which at least a portion of the target space is captured at the second time point; The information acquisition unit further acquires an infrared image in which at least a portion of the target space is captured at the first time point, The prediction device according to claim 1 , wherein the index acquisition unit further inputs the infrared image acquired by the information acquisition unit into the prediction model.

4. At least a part of the outline of the target space is formed by a target part that is part of the real estate; the example data includes second learning temperature information indicating a temperature of a surface of the target portion facing the target space at the second time point; the information acquisition unit further acquires second temperature information indicating a temperature of a surface of the target portion on the target space side at the first time point; The prediction device according to claim 1 , wherein the index acquisition unit further inputs the second temperature information acquired by the information acquisition unit into the prediction model.

5. There are multiple spaces other than the target space within the real estate, The example data includes: Third learning temperature information indicating temperatures at the second time point in at least some of the spaces; fourth learning temperature information indicating the outside air temperature at the second time point; a first learning contribution indicating the degree of influence that the temperature at the second time point in the at least some space has on the temperature of the target space; a second contribution indicating a degree of influence of the outside air temperature at the second time point on the temperature of the target space, the information acquisition unit further acquires third temperature information indicating temperatures at the first time point in at least some of the spaces among the plurality of spaces, and fourth temperature information indicating an outside air temperature at the first time point; a temperature contribution calculation unit that calculates a first contribution indicating the degree of influence that the temperature in the at least some space at the first time point has on the temperature of the target space, and a second contribution indicating the degree of influence that the outside air temperature at the first time point has on the temperature of the target space, based on the first temperature information, the third temperature information, and the fourth temperature information acquired by the information acquisition unit; The prediction device according to claim 1 , wherein the index acquisition unit further inputs the first contribution and the second contribution calculated by the temperature contribution calculation unit into the prediction model.

6. At least a part of the outline of the target space is formed by a target part that is part of the real estate; The correct answer data includes an actual measurement value of an insulation index indicating the insulation performance of the target portion at a predetermined time after the second time point in the target portion, The prediction device according to claim 5 , wherein the index acquisition unit further acquires, from the prediction model, a predicted value of the insulation index in the target location at a predetermined time period after the first time point.

7. The prediction device according to claim 5 , further comprising a display control unit that causes at least one of the first contribution degree and the second contribution degree calculated by the temperature contribution degree calculation unit to be displayed on a display unit.

8. There are multiple spaces other than the target space within the real estate, At least a part of the outline of the target space is formed by a target part that is part of the real estate; The example data includes: Second learning humidity information indicating humidity at the second time point in at least some of the spaces among the plurality of spaces; third learning humidity information indicating the humidity outside the real estate at the second time point; a third learning contribution indicating the degree of influence that the humidity at the second time point in the at least part of the space has on the humidity of the target space; and a fourth learning contribution indicating the degree of influence of the humidity outside the real estate at the second time point on the humidity in the target space, The correct answer data includes an actual measurement value of an airtightness index indicating the airtightness performance of the target portion at a predetermined time after the second time point in the target portion, the information acquisition unit further acquires second humidity information indicating humidity at the first time point in at least some of the spaces, and third humidity information indicating humidity outside the real estate at the first time point; a humidity contribution calculation unit that calculates a third contribution indicating the degree of influence that the humidity in the at least some space at the first time point has on the humidity in the target space, and a fourth contribution indicating the degree of influence that the humidity outside the real estate at the first time point has on the humidity in the target space, based on the first humidity information, the second humidity information, and the third humidity information acquired by the information acquisition unit; The prediction device described in claim 1, wherein the index acquisition unit further inputs each of the third contribution rate and the fourth contribution rate calculated by the humidity contribution rate calculation unit into the prediction model, thereby further acquiring a predicted value of the airtightness index for the target area at a predetermined period after the first point in time.

9. A prediction program for causing a computer to function as the prediction device according to claim 1, the prediction program causing a computer to function as the information acquisition unit and the index acquisition unit.

10. A prediction system for predicting the amount of energy consumed by air conditioning equipment installed in a target space within real estate, an information acquisition unit that acquires first temperature information indicating a temperature of the target space at a first time point and first humidity information indicating a humidity of the target space at the first time point; a prediction system comprising: an index acquisition unit that inputs the first temperature information and the first humidity information acquired by the information acquisition unit into a machine-learned prediction model using teacher data in which first learning temperature information indicating the temperature of the target space at a second time point different from the first time point and first learning humidity information indicating the humidity of the target space at the second time point are used as example data, and a learning energy index indicating the amount of energy consumed at a specified time period after the second time point is used as correct answer data, and acquires an energy index indicating the prediction result of the amount of energy consumed at a specified time period after the first time point.

11. A prediction method executed by one or more computers for predicting an amount of energy consumption of an air conditioning device installed in a target space in real estate, comprising: an information acquisition step of acquiring first temperature information indicating a temperature of a target space in the real estate at a first time point and first humidity information indicating humidity of the target space at the first time point; a prediction method including: an index acquisition step of inputting the first temperature information and the first humidity information acquired in the information acquisition step into a machine-learned prediction model using teacher data in which first learning temperature information indicating the temperature of the target space at a second time point different from the first time point and first learning humidity information indicating the humidity of the target space at the second time point are used as example data, and a learning energy index indicating the amount of energy consumed at a specified time period after the second time point is used as correct answer data, and acquiring an energy index indicating the predicted result of the amount of energy consumed at a specified time period after the first time point.

Images