Information providing device, information providing method and program

Abstract

To provide an information providing device, an information providing method, and a program capable of appropriately and accurately predicting an electric power consumption pattern and providing useful information based on the predicted pattern.SOLUTION: A cloud server 1, which is a kind of an information providing device, includes: a vector generation unit 112 that generates a plurality of electric power amount vectors whose elements are electric power consumption amounts consumed by electric apparatuses; an element decomposition unit 113 that calculates, for each of a plurality of target periods, a plurality of weighting factors that corresponds to respective element vectors when decomposing the plurality of electric power amount vectors into the plurality of element vectors; and a pattern identifying unit 114 that identifies an electric power consumption pattern for each of the plurality of target periods, by clustering the plurality of target periods into a plurality of electric power consumption patterns using a weighting factor vector whose elements are the plurality of weighting factors that corresponds to the respective target periods.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present disclosure relates to an information providing device, an information providing method, and a program. [Background technology]

[0002] A lifestyle pattern estimation device has been proposed that includes clustering means for generating feature vectors for each statistical target user for each target period based on power consumption data within a specific time period and classifying the feature vectors into clusters; lifestyle behavior estimation means for estimating occurrence times of lifestyle behaviors based on the power consumption data of the statistical target user; labeling means for labeling the classified clusters with lifestyle behaviors and occurrence times based on the feature vectors belonging to the clusters and the occurrence times of the lifestyle behaviors estimated for the feature vectors; and lifestyle pattern estimation means for determining a cluster corresponding to the power consumption data of a specific user from the classified clusters and estimating the lifestyle pattern of the specific user based on the lifestyle behaviors and occurrence times labeled for the determined cluster (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2015-197741 Summary of the Invention [Problem to be solved by the invention]

[0004] Incidentally, a lifestyle pattern estimation device such as that described in Patent Document 1 is required to have a function of appropriately and accurately estimating the lifestyle patterns of users to be statistically analyzed and providing useful information based on the estimated lifestyle patterns.

[0005] The present disclosure has been made in consideration of the above-mentioned reasons, and aims to provide an information providing device, an information providing method, and a program that can appropriately and accurately predict power consumption patterns and provide useful information based on the predicted patterns. [Means for solving the problem]

[0006] In order to achieve the above object, an information providing device according to the present disclosure includes: a vector generation unit that generates a plurality of power amount vectors, each of which corresponds to a plurality of target periods for which at least one power consumption pattern is to be identified, and each of which has as elements an amount of power consumption consumed by at least one electrical device in each of a plurality of time periods that make up the target periods; an element decomposition unit that calculates, for each of the plurality of target periods, a plurality of weighting coefficients corresponding to each of the plurality of element vectors obtained by decomposing the plurality of power amount vectors into a plurality of element vectors; a pattern identification unit that identifies the power consumption pattern for each of the plurality of target periods by clustering the plurality of target periods into a plurality of power consumption patterns using a weight coefficient vector having the plurality of weight coefficients corresponding to the plurality of target periods as elements; and a notification unit that generates notification information for notifying a user of content related to the power consumption pattern based on the power consumption pattern for each of the plurality of target periods and transmits the notification information to the terminal device. [Effects of the Invention]

[0007] According to the present disclosure, an element decomposition unit calculates, for each of a plurality of target periods, a plurality of weighting factors corresponding to each of the plurality of element vectors obtained by decomposing a plurality of power energy vectors into a plurality of element vectors. Then, a pattern identification unit identifies a power consumption pattern for each of the plurality of target periods by clustering the plurality of target periods into a plurality of power consumption patterns using a weighting factor vector whose elements are the plurality of weighting factors corresponding to each of the plurality of target periods. This allows the power consumption pattern for each of the plurality of past target periods to be expressed using a finite number of element vectors and their corresponding weighting factors, thereby highlighting the characteristics of the power consumption pattern for each of the plurality of past target periods and thereby enabling appropriate and accurate prediction of the power consumption pattern for each of the plurality of target periods. This reduces the provision of notification information to users based on erroneously predicted power consumption patterns, thereby providing users with useful notification information. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a schematic configuration diagram of an information providing system according to a first embodiment of the present disclosure; [Figure 2] FIG. 1 is a block diagram showing a hardware configuration of a cloud server according to a first embodiment. [Figure 3] FIG. 1 is a block diagram showing a functional configuration of a cloud server according to a first embodiment. [Figure 4] FIG. 1 is a diagram showing an example of information stored in an electric energy storage unit according to the first embodiment; [Figure 5] FIG. 1 is a diagram showing an example of information stored in a processing object storage unit according to the first embodiment; [Figure 6] FIG. 1 is a diagram showing an example of information stored in an element information storage unit according to the first embodiment; [Figure 7] FIG. 10 is a diagram showing an example of information stored in a pattern storage unit according to the first embodiment; [Figure 8] An explanatory diagram of the operation of the element decomposition unit according to the first embodiment. [Figure 9] An explanatory diagram of the operation of a pattern identification unit according to the first embodiment. [Figure 10] 10 is a flowchart showing an example of the flow of information providing processing executed by the cloud server according to the first embodiment. [Figure 11] 10 is a flowchart showing an example of the flow of information providing processing executed by the cloud server according to the first embodiment. [Figure 12] An explanatory diagram of the operation of a pattern identification unit according to the first embodiment. [Figure 13] FIG. 1A is a diagram showing an example of a notification image displayed on a display unit of a terminal device carried by an administrator according to the first embodiment; FIG. 1B is a diagram showing an example of a notification image displayed on a display unit of a terminal device carried by a user of an electrical device according to the first embodiment; [Figure 14] FIG. 10 is a block diagram showing a functional configuration of a cloud server according to a second embodiment of the present disclosure. [Figure 15] FIG. 10 is a diagram showing an example of information stored in a processing object storage unit according to the second embodiment; [Figure 16] 10 is a flowchart showing an example of the flow of information providing processing executed by a cloud server according to a second embodiment. [Figure 17] 10 is a flowchart showing an example of the flow of information providing processing executed by a cloud server according to a second embodiment. [Figure 18] FIG. 10 is a block diagram showing the functional configuration of a cloud server according to a modified example. [Figure 19] FIG. 1A is a diagram showing an example of information stored in a weather information storage unit according to a modified example, and FIG. 1B is a diagram showing an example of information stored in a processing object storage unit according to a modified example. [Figure 20] FIG. 10 is a diagram showing an example of information stored in an element information storage unit according to a modified example. [Figure 21] FIG. 10 is a diagram showing an example of information stored in a pattern storage unit according to a modified example. [Figure 22] 10 is a flowchart showing an example of the flow of an information providing process executed by a cloud server according to a modified example. [Figure 23] FIG. 10 is a block diagram showing the functional configuration of a cloud server according to a modified example. [Figure 24] 10 is a flowchart showing an example of the flow of an information providing process executed by a cloud server according to a modified example. DETAILED DESCRIPTION OF THE INVENTION

[0009] (Embodiment 1) An information providing device according to each embodiment of the present disclosure will be described below with reference to the drawings. The information providing device according to the present embodiment includes: a vector generating unit that generates a plurality of power amount vectors, each of which corresponds to a plurality of target periods for which at least one power consumption pattern is to be identified, and whose elements are the amount of power consumed by at least one electrical device in each of a plurality of time periods constituting the target periods; an element decomposition unit that calculates, for each of the plurality of target periods, a plurality of weighting coefficients corresponding to each of the plurality of element vectors obtained by decomposing the plurality of power amount vectors into a plurality of element vectors; a pattern identifying unit that identifies a power consumption pattern for each of the plurality of target periods by clustering the plurality of target periods into a plurality of power consumption patterns using a weighting coefficient vector whose elements are the plurality of weighting coefficients corresponding to each of the plurality of target periods; and a notification unit that generates notification information notifying a user of content related to the power consumption pattern, based on the power consumption pattern for each of the plurality of target periods, and transmits the notification information to a terminal device.

[0010] The information providing device according to this embodiment notifies notification information based on the power consumption patterns of a house in which multiple types of electrical appliances 4 are installed, as shown in FIG. 1 . Here, a power grid PS supplies power to multiple electrical appliances 4 (e.g., a lighting fixture 4A, an air conditioner 4B, an induction cooker 4C, a dishwasher 4D, and a microwave oven 4E) via a distribution board 21 installed in the house. The multiple electrical appliances 4 are connected to branch lines L1, L2, L3, L4, and L5 extending from the distribution board 21. Also installed in the house is a watt-hour meter 22 that measures the amount of power supplied to each of the multiple electrical appliances 4. The watt-hour meter 22 and the multiple electrical appliances 4 each communicate with the control device 2 via a local network NW2. The local network NW2 is a wireless local area network (LAN) or a wired LAN. Also connected to this local network NW2 is a broadband router (hereinafter referred to as "BBR") 6 that is connected to a wide area network NW1 such as the Internet. In addition, the cloud server 1 functioning as an information providing device, a terminal device 51 carried by a user of the electrical device 4, and a terminal device 52 carried by an administrator supporting the user can communicate with each other via the wide area network NW1. In addition, the control device 2 communicates with the cloud server 1 via the BBR 6 and the wide area network NW1.

[0011] The watt-hour meter 22 measures the amount of power consumed by each of the plurality of electrical devices 4 based on the value of the current flowing through each of the branch lines L1, L2, L3, L4, and L5, generates power information indicating the measured amount of power consumption, and transmits the information to the control device 2. The control device 2 acquires the power information transmitted from the watt-hour meter 22, and transmits the acquired power information to the cloud server 1 via the BBR6 and the wide area network NW1.

[0012] Terminal devices 51 and 52 are, for example, smartphones, and include a display unit and an input unit. The display unit is a liquid crystal display, an organic EL (Electro-Luminescence) display, or the like, and displays a notification image on the display unit based on information acquired from cloud server 1. The input unit is, for example, a transparent touchpad placed over the display unit, and receives various operational information in response to operations by a user or administrator.

[0013] The cloud server 1 is, for example, a computer used as a server, and as shown in FIG. 2, includes a CPU (Central Processing Unit) 101, a main memory 102, an auxiliary memory 103, a wide area communication unit 106 connected to a wide area network NW1, a clock unit 108, and a bus 109 interconnecting these components. The main memory 102 is a volatile memory used as a work area for the CPU 101. The auxiliary memory 103 is a nonvolatile memory such as a magnetic disk or semiconductor flash memory, and stores programs for implementing various functions of the cloud server 1. The wide area communication unit 106 communicates with the power measurement device 22 via the wide area network NW1, BBR6, and local network NW2. The clock unit 108 functions as an RTC (Real Time Clock) that keeps track of the current date and time.

[0014] The CPU 101 loads the programs stored in the auxiliary storage unit 103 into the main storage unit 102 and executes them, thereby functioning as an energy amount obtaining unit 111, a vector generating unit 112, an element decomposition unit 113, a pattern identifying unit 114, a number-of-patterns determining unit 115, a notification determining unit 116, and a notification unit 117, as shown in Fig. 3. The auxiliary storage unit 103 shown in Fig. 2 also includes an energy amount storage unit 131, a vector storage unit 132, a processing target storage unit 133, an element information storage unit 134, a pattern storage unit 135, and a notification information storage unit 136, as shown in Fig. 3. The energy amount storage unit 131 stores energy amount information indicating the amount of energy supplied to the electrical appliances 4 via the branch lines L1, L2, L3, L4, and L5 in association with a combination of time information, branch line information, appliance type information, and house identification information, as shown in Fig. 4, for example. Here, the time information indicates the start of the time period when the amount of power indicated by the power amount information was measured, the branch line information indicates identification information for the branch lines L1, L2, L3, L4, and L5, the device type information indicates the type of electrical device 4 connected to the branch lines L1, L2, L3, L4, and L5, and the house identification information is information for identifying the house in which the electrical device 4 is installed.

[0015] 5, the processing target storage unit 133 stores a combination of target period group information indicating a period to which a plurality of target periods for which at least one power consumption pattern is to be identified belong, branch line information identifying the branch lines L1, L2, L3, L4, and L5 for which power consumption amounts are to be acquired, and home identification information identifying the homes in which the branch lines L1, L2, L3, L4, and L5 are installed, in association with process identification information for identifying a process for identifying a power consumption pattern. Here, the target period is set to, for example, a period of one day (24 hours). The period indicated by the target period group information is set to, for example, a continuous period prior to the present time.

[0016] Returning to FIG. 3 , the vector storage unit 132 stores power amount vector information indicating a plurality of power amount vectors, each of which corresponds to a plurality of predetermined target periods and has as its elements the amount of power consumed by the electrical appliances 4 in each of a plurality of time slots constituting the target periods. Here, the plurality of target periods are included in the target period indicated by the target period information stored in the processing target storage unit 133. The target period is set to, for example, one day. In this case, the vector storage unit 132 stores the power amount vector information corresponding to the target period corresponding to the target period in association with the target period information indicating the target period. The plurality of time slots can be, for example, 24 time slots, each one hour long, obtained by dividing a day into one-hour increments.

[0017] As shown in FIG. 6, for example, the element information storage unit 134 stores element vector information indicating multiple element vectors for each target period and weight coefficient information indicating weight coefficients corresponding to each element vector, in association with the processing identification information described above, target period information indicating the target period, and element identification information identifying each element vector.

[0018] 7, the pattern storage unit 135 stores a combination of candidate pattern identification information for identifying candidate power consumption patterns to which the element vectors of each target period belong, average weighting coefficient information, and probability information indicating the probability that each candidate pattern belongs, in association with the target period information and processing identification information. Here, the pattern identification information is information for identifying candidate power consumption patterns to which the element vectors of each target period belong when the element vectors are clustered into multiple power consumption patterns. Furthermore, the average weighting coefficient information indicates the average value of weighting coefficients corresponding to the multiple element vectors in each power consumption pattern.

[0019] 3, the notification information storage unit 136 stores multiple types of message information included in the notification information transmitted to the terminal devices 51 and 52. The terminal identification information storage unit 137 stores terminal identification information for identifying the terminal devices 51 and 52 to which the notification information is transmitted. The terminal identification information may include, for example, the IP addresses and port numbers of the terminal devices 51 and 52.

[0020] The power amount acquiring unit 111 acquires power amount information indicating the amount of power consumed by each of the plurality of electrical appliances 4 from the control device 2. When the power amount acquiring unit 111 determines that a preset time for acquiring power amount information has arrived based on the date and time measured by the clock unit 108, the power amount acquiring unit 111 transmits power amount information request information to the control device 2 requesting the control device 2 to transmit the power amount information, thereby acquiring the power amount information transmitted from the control device 2. Then, the power amount acquiring unit 111 stores the acquired power amount information in the power amount storage unit 131.

[0021] The vector generation unit 112 generates an electric energy vector corresponding to each of a plurality of target periods for identifying at least one of the above-mentioned power consumption patterns, using the electric energy information stored in the electric energy storage unit 131. Here, the electric energy vector is a vector whose elements are the amounts of power consumed by the electrical appliances 4 in each of a plurality of time periods constituting the above-mentioned target period. The vector generation unit 112 stores electric energy vector information indicating the generated electric energy vector in the vector storage unit 132 in association with target period information indicating the corresponding target period.

[0022] The element decomposition unit 113 calculates, for each of a plurality of target periods, a plurality of weighting factors corresponding to each of the plurality of element vectors obtained by decomposing a plurality of power energy vectors into a plurality of element vectors. The element decomposition unit 113 calculates, for each of the plurality of target periods, a plurality of element vectors and a plurality of weighting factors corresponding to the plurality of element vectors, for example, based on the nonnegative matrix factorization (NMF) method. Here, the element decomposition unit 113 first calculates the plurality of element vectors and weighting factors for each target period, for example, based on the NMF method, with the number of element vectors set as a predetermined initial number of element vectors. Here, the initial number of element vectors is set to, for example, two. Next, the element decomposition unit 113 calculates the difference for each element between the power energy vector generated for each target period based on the calculated element vectors and weighting factors and the power energy vector stored in the vector storage unit 132, thereby calculating the error of the power energy vector. For example, the sum of the absolute values ​​of the differences calculated for each element, or the standard deviation or variance of the differences for all elements can be used as this error. Specifically, for example, as shown on the left side of Fig. 8, the element decomposition unit 113 calculates four element vectors Vc1, Vc2, Vc3, and Vc4 and their corresponding weighting coefficients "0.31," "0.32," "0.30," and "0.1" for a certain target period. In this case, the element decomposition unit 113 generates an energy vector VP2 indicated by a dashed line on the right side of Fig. 8 based on the calculated four element vectors Vc1, Vc2, Vc3, and Vc4 and their corresponding weighting coefficients "0.31," "0.32," "0.30," and "0.1." The element decomposition unit 113 then calculates the error of the energy vector by calculating the difference for each element between the energy vector VP2 and the energy vector VP1 indicated by a solid line on the right side of Fig. 8 and stored in the vector storage unit 132. Next, the element decomposition unit 113 calculates the average, median or maximum value of the error of the power vector for all target periods included in the aforementioned target period group, and determines whether the calculated average, median or maximum value of the error is greater than or equal to a predetermined error threshold value.If the calculated average, median, or maximum value of the error is equal to or greater than the error threshold, the element decomposition unit 113 increases the number of element vectors by a preset number (for example, 1), and then calculates the power amount vector for each of the plurality of target periods again, along with the plurality of element vectors and the plurality of weighting coefficients corresponding thereto.On the other hand, if the calculated average, median, or maximum value of the error is less than the error threshold, the element decomposition unit 113 stores element vector information indicating the plurality of calculated element vectors and weighting coefficient information indicating the plurality of weighting coefficients corresponding thereto in the element information storage unit 134 in association with the element identification information, target period information, and processing identification information.

[0023] Returning to FIG. 3 , the pattern identification unit 114 identifies a power consumption pattern for each of the target periods by clustering the target periods into multiple power consumption patterns using a weighting coefficient vector having multiple weighting coefficients corresponding to each of the target periods as elements. The pattern identification unit 114 identifies at least one power consumption pattern for each of the target periods by clustering the weighting coefficient vector corresponding to each of the target periods into multiple power consumption patterns using a method using the K-means algorithm or a Gaussian Mixture Model (GMM). Specifically, assuming that a clustering method that evaluates the probability of belonging to each cluster, such as the GMM, is used, the pattern identification unit 114 identifies the power consumption pattern with the highest probability of belonging to each of the target periods, as shown in the lower part of FIG. 9 , when the weighting coefficients corresponding to four element vectors Vc1, Vc2, Vc3, and Vc4 fluctuate over time within the target period, as shown in the upper part of FIG. 9 . Furthermore, even when a clustering method that uniquely determines which cluster a pattern belongs to, such as the K-means algorithm, is used, the power consumption pattern is similarly identified by setting the probability of belonging to that cluster to 1 (100%). Next, the pattern identification unit 114 calculates, for each of the multiple power consumption patterns, an average power energy vector of the power energy vectors for the target period corresponding to at least one weighting coefficient vector belonging to each power consumption pattern.The pattern identification unit 114 then compares the magnitudes of the difference vectors between the multiple average power energy vectors, and if there are multiple average power energy vectors whose magnitudes are equal to or less than a predetermined difference threshold, performs a pattern integration process to integrate the multiple power consumption patterns corresponding to the multiple average power energy vectors into a single power consumption pattern.Then, for each of the multiple power consumption patterns after the integration process, the pattern identification unit 114 calculates an average weighting coefficient vector whose elements are the average values ​​of the elements constituting at least one weighting coefficient vector, and a corresponding probability.Then, the pattern identification unit 114 stores average weighting coefficient vector information indicating the average weighting coefficient vector of the power consumption pattern in each target period and corresponding probability information indicating the corresponding probability in the pattern storage unit 135 in association with the pattern identification information, target period information, and processing identification information.

[0024] 3, the notification determination unit 116 determines whether or not to notify the user of notification information related to the power consumption pattern, based on the power consumption pattern for each of the multiple target periods stored in the pattern storage unit 135. The notification determination unit 116 determines whether at least one power consumption pattern identified based on the multiple power amount vectors corresponding to the most recent first time point and each of the multiple target periods earlier than the first time point differs from at least one power consumption pattern identified based on the multiple power amount vectors corresponding to each of the multiple target periods earlier than a second time point that is a preset set period earlier than the first time point. Furthermore, the notification determination unit 116 refers to the pattern identification information stored in the pattern storage unit 135 and determines whether the power consumption pattern at the most recent first time point has transitioned to a power consumption pattern different from the power consumption pattern earlier than the first time point.

[0025] The notification unit 117 generates notification information notifying the user of content related to the power consumption pattern based on the power consumption pattern for each of the plurality of target periods, and transmits the notification information to the terminal devices. Specifically, when the notification determination unit 116 determines at the first time point that at least one power consumption pattern for each of the two types of periods is different from each other, the notification unit 117 generates notification information notifying that the user's lifestyle pattern has changed, and transmits the notification information to the terminal devices 51 and 52. Furthermore, when the notification determination unit 116 determines that the power consumption pattern at the most recent first time point has transitioned to a power consumption pattern different from the power consumption pattern prior to the first time point, the notification unit 117 generates notification information related to a setting change of the electric appliance 4 due to the change in the user's lifestyle pattern, and transmits the notification information to the terminal devices 51 and 52.

[0026] Next, the information providing process executed by the cloud server 1 according to this embodiment will be described with reference to Fig. 10 to Fig. 13. This information providing process is started when a program for executing the information providing process is started in the cloud server 1. In parallel with this information providing process, the power amount obtaining unit 111 obtains the power amount information transmitted from the control device 2 by transmitting power amount information request information to the control device 2 requesting the control device 2 to transmit power amount information each time a power consumption information obtaining period arrives, and stores the obtained power amount information in the power amount storage unit 131.

[0027] First, as shown in FIG. 10, the vector generation unit 112 determines whether a pattern confirmation time has arrived, at which a preset power consumption pattern is to be confirmed (step S101). As long as the vector generation unit 112 determines that the pattern confirmation time has not yet arrived (step S101: No), the vector generation unit 112 repeatedly executes the process of step S101. On the other hand, when the vector generation unit 112 determines that the pattern confirmation time has arrived (step S101: Yes), the vector generation unit 112 identifies the power energy information to be used from the power energy information stored in the power energy storage unit 131 (step S102). Here, the vector generation unit 112 identifies a target period group from the target period group stored in the processing target storage unit 133, and identifies power energy information indicating the amount of power consumption in each target period included in the identified target period group. Next, the vector generation unit 112 uses the identified power energy information to generate a power energy vector corresponding to each of the target periods included in the target period (step S103). The vector generating unit 112 stores the generated power amount vector information in the vector storage unit 132 in association with target period information indicating the corresponding target period.

[0028] Next, the element decomposition unit 113 calculates, for each of the multiple target periods, multiple weighting factors corresponding to each of the multiple element vectors obtained by decomposing the multiple energy vectors stored in the vector storage unit 132 into multiple element vectors (step S104). After that, the element decomposition unit 113 calculates the difference for each element between the energy vector generated based on the calculated element vectors and weighting factors and the energy vector stored in the vector storage unit 132 for each target period, thereby calculating the error of the energy vector (step S105). Next, the element decomposition unit 113 calculates the average, median, or maximum value of the error of the energy vector for all target periods included in the target period, and determines whether the calculated value E is equal to or greater than a predetermined error threshold Eth (step S106). Here, if the element decomposition unit 113 determines that the calculated value E is equal to or greater than the error threshold Eth (step S106: Yes), it increases the number of element vectors by a preset number (for example, one) (step S107) and then executes the processing of step S104 again.

[0029] On the other hand, if the element decomposition unit 113 determines that the calculated value E is less than the error threshold Eth (step S106: No), it stores element vector information indicating the calculated multiple element vectors and weight coefficient information indicating the corresponding multiple weight coefficients in the element information storage unit 134 in association with the element identification information, target period information, and process identification information. Then, the pattern identification unit 114 generates a weight coefficient vector having multiple weight coefficients corresponding to each of the multiple target periods as elements (step S108). Next, the pattern identification unit 114 clusters the multiple target periods into multiple power consumption patterns using the generated multiple weight coefficient vectors corresponding to the multiple target periods (step S109). As a result, the pattern identification unit 114 identifies at least one power consumption pattern for each of the multiple target periods.

[0030] Then, for each of the plurality of power consumption patterns, the pattern identification unit 114 calculates an average power energy vector of the power energy vectors for the target period corresponding to at least one weight coefficient vector belonging to the power consumption pattern, and calculates a difference vector between the calculated plurality of average power energy vectors (step S110). Then, the pattern identification unit 114 determines whether or not there are a plurality of average power energy vectors whose magnitude dP of the difference vector is equal to or less than a predetermined difference threshold dPth (step S111). If the pattern identification unit 114 determines that there is no average power energy vector whose magnitude |dVp| of the difference vector is equal to or less than the difference threshold |dVpth| (step S111: No), the process of step S113 described below is executed. On the other hand, if the pattern identification unit 114 determines that there are a plurality of average power energy vectors whose magnitude dP of the difference vector is equal to or less than a predetermined difference threshold dPth (step S111: Yes), the pattern identification unit 114 executes an integration process to integrate the power consumption patterns corresponding to the plurality of average power energy vectors into a single power consumption pattern (step S112). In this integration process, the pattern identification unit 114 identifies a weight coefficient vector for each of the target periods belonging to the multiple power consumption patterns to be integrated. Next, the pattern identification unit 114 calculates a new average weight coefficient for each of the multiple power consumption patterns. Then, the pattern identification unit 114 stores average weight coefficient vector information indicating an average weight coefficient vector whose elements are average weight coefficients corresponding to the power consumption pattern to which the target period belongs, and probability information indicating the probability of the power consumption pattern being matched, in the pattern storage unit 135, in association with the pattern identification information, target period information, and process identification information (step S113).

[0031] Next, the pattern number determination unit 115 determines whether the number Npa of the multiple power consumption patterns is equal to or greater than the predetermined pattern number threshold Npath (step S114). If the pattern number determination unit 115 determines that the number Npa of the multiple power consumption patterns is equal to or greater than the pattern number threshold Npath (step S114: Yes), generation and transmission of notification information is avoided, and the process of step S101 is executed again. On the other hand, assume that the pattern number determination unit 115 determines that the number Npa of the multiple power consumption patterns is less than the pattern number threshold Npath (step S114: No). In this case, as shown in FIG. 11 , the notification determination unit 116 determines whether a power consumption pattern identified based on a plurality of power amount vectors corresponding to a first time point and a plurality of target periods earlier than the first time point stored in the pattern storage unit 135 differs from a power consumption pattern identified based on a plurality of power amount vectors corresponding to a plurality of target periods earlier than a second time point that is the aforementioned set period earlier than the first time point (step S115). 12, for example, assume that the power consumption patterns identified based on the plurality of power amount vectors corresponding to the plurality of target periods prior to the second time point, which is the aforementioned set period prior to the first time point, are only the power consumption patterns identified by pattern identification information “1” and “2.” In this case, assume that the power consumption patterns identified based on the plurality of power amount vectors corresponding to the first time point and the plurality of target periods prior to the first time point, as shown in the power consumption pattern corresponding to the process identification information 0002, are the power consumption patterns identified by pattern identification information “1,” “2,” and “3.” In this case, the notification determination unit 116 determines that the power consumption patterns are different because the presence or absence of the power consumption pattern identified by pattern identification information “3” differs. If the notification determination unit 116 determines that the power consumption patterns are the same (step S115: No), the process proceeds to step S117, which will be described later.

[0032] On the other hand, if the notification determination unit 116 determines that the power consumption pattern is different (step S115: Yes), the notification unit 117 generates notification information notifying the user of a change in their lifestyle pattern using the message information stored in the notification information storage unit 136, and transmits the notification information to the terminal devices 51 and 52 (step S116). The notification information transmitted to the terminal device 51 includes the most recent power energy vector information, and element vector information and weighting coefficient information corresponding to the power consumption pattern identified based on the multiple power energy vectors corresponding to the first time point and multiple target periods prior to the first time point. Then, as shown in FIG. 13(A), for example, the terminal device 51 possessed by the administrator generates a power energy vector based on the element vector information and the weighting coefficient information, and displays a notification image GA11 on the display unit that includes an image representing the generated power energy vector as a "normal" power energy vector and an image representing the most recent power energy vector. Meanwhile, the terminal device 52 carried by the user of the electrical appliance 4 displays, on the display unit, a notification image GA12 for notifying the user of a change in lifestyle pattern, as shown in FIG. 13(B), for example.

[0033] 11 , the notification determination unit 116 then refers to the pattern identification information stored in the pattern storage unit 135 and determines whether the power consumption pattern at the most recent first time point has transitioned to a power consumption pattern different from a power consumption pattern in the past than the first time point (step S117). If the notification determination unit 116 determines that the power consumption pattern before the first time point has remained constant (step S117: No), the process of step S101 is executed again. On the other hand, if the notification determination unit 116 determines that the power consumption pattern at the most recent first time point has transitioned to a power consumption pattern different from a pattern in the past than the first time point (step S117: Yes), the notification determination unit 116 generates notification information regarding a change in the settings of the electric appliance 4 due to a change in the user's lifestyle pattern, using the message information stored in the notification information storage unit 136, and transmits the notification information to the terminal devices 51 and 52 (step S118).

[0034] As described above, in the cloud server 1 according to the present embodiment, the element decomposition unit 113 calculates, for each of a plurality of target periods, a plurality of weighting factors corresponding to each of the plurality of element vectors obtained by decomposing a plurality of power energy vectors into a plurality of element vectors. The pattern identification unit 114 then clusters the plurality of target periods into a plurality of power consumption patterns using a weighting factor vector whose elements are the plurality of weighting factors corresponding to each of the plurality of target periods, thereby identifying the power consumption pattern for each of the plurality of target periods. This allows the power consumption pattern for each of the plurality of past target periods to be expressed using a finite number of element vectors and their corresponding weighting factors, thereby highlighting the characteristics of the power consumption pattern for each of the plurality of past target periods. This allows the power consumption pattern for each of the plurality of target periods to be predicted appropriately and accurately. This reduces the provision of notification information based on erroneously predicted power consumption patterns to users, thereby providing notification information that is useful to users.

[0035] Furthermore, when the pattern identification unit 114 according to this embodiment identifies multiple power consumption patterns, it calculates the average power consumption amount for each of the multiple power consumption patterns in multiple time periods. If the average, median, or maximum value of the difference in the average power consumption amounts for each of the multiple time periods between two power consumption patterns selected from the calculated multiple power consumption patterns is equal to or less than a preset difference threshold, the pattern identification unit 114 executes a pattern integration process to integrate the two selected power consumption patterns into a single power consumption pattern. This reduces the number of power consumption patterns, thereby preventing the provision of more notification information than necessary to users or administrators.

[0036] Furthermore, the vector generation unit 112 according to this embodiment generates a plurality of power amount vectors whose elements are the amount of power consumed in each of a plurality of time periods during a target period for each of the plurality of electrical appliances 4. This makes it possible to provide notification information to a user or an administrator based on a combination of power consumption patterns of the plurality of electrical appliances 4, thereby enabling more detailed classification of user behavior. Therefore, it is possible to provide appropriate notification information that matches the user's behavior pattern.

[0037] Furthermore, the notification determination unit 116 according to the present embodiment determines whether a power consumption pattern identified based on a plurality of power vectors corresponding to a first time point and a plurality of target periods in the past before the first time point differs from a power consumption pattern identified based on a plurality of power vectors corresponding to a plurality of target periods in the past before a second time point that is a predetermined set period in the past before the first time point. If the notification unit 117 determines that these two power consumption patterns differ from each other, it transmits notification information to the terminal devices 51 and 52 to notify them that the lifestyle pattern of the user of the electrical appliance 4 has changed. Here, the notification unit 117 transmits the notification information to the terminal devices 51 and 52 identified by the terminal identification information stored in the terminal identification information storage unit 137. This makes it possible to detect a change in the lifestyle pattern of the user of the electrical appliance 4, thereby enabling, for example, monitoring of the user based on the user's behavioral patterns.

[0038] Furthermore, the pattern number determination unit 115 according to this embodiment determines whether the number of patterns of the multiple power consumption patterns obtained when multiple target periods are clustered into multiple power consumption patterns is equal to or greater than a preset pattern number threshold. Then, the notification unit 117 avoids transmitting to the terminal devices 51 and 52 notification information based on the power consumption patterns of the electrical appliances 4 for which the number of patterns is determined to be equal to or greater than the aforementioned pattern number threshold. This allows the power consumption patterns of the electrical appliances 4 used to generate the notification information to be selected, and makes it possible to prevent excessive notification to the terminal devices 51 and 52 of notification information based on the power consumption patterns of the electrical appliances 4 for which the number of patterns is equal to or greater than the aforementioned pattern number. Therefore, it is possible to provide the notification information appropriately to the user.

[0039] (Embodiment 2) The information providing device according to this embodiment differs from the first embodiment in that it includes an operation timing estimation unit that identifies an element vector with a maximum weighting coefficient for each of at least one power consumption pattern identified for a plurality of predetermined target periods, and that identifies a predicted operation timing for at least one electrical device based on the identified element vector. Furthermore, the electrical device according to this embodiment has a function of starting or stopping operation based on the acquired control information when it acquires control information from the cloud server 1.

[0040] The hardware configuration of the cloud server according to this embodiment is the same as the hardware configuration described in the first embodiment. In the description of this embodiment, the same hardware configuration as that of the first embodiment will be described using the same reference numerals as those shown in FIG. 2 as appropriate. As shown in FIG. 14, the CPU 101 of the cloud server 2001 according to this embodiment reads out a program stored in the auxiliary storage unit 103 into the main storage unit 102 and executes it, thereby functioning as an energy acquisition unit 111, a vector generation unit 112, a decomposition unit 113, a pattern identification unit 114, a number of patterns determination unit 115, a notification unit 2117, an operation time estimation unit 2118, a notification time identification unit 2119, a similar operation time determination unit 2120, a control information generation unit 2121, and a control information transmission unit 2122. Note that in FIG. 14, the same components as those in the first embodiment are denoted by the same reference numerals as those in FIG. 3. 14, the auxiliary storage unit 103 includes an electric energy storage unit 131, a vector storage unit 132, a processing target storage unit 133, an element information storage unit 134, a pattern storage unit 135, a notification information storage unit 2136, a terminal identification information storage unit 137, and an operation period storage unit 2138. As shown in Fig. 15, for example, the processing target storage unit 133 acquires the amount of power consumed from each of a plurality of branch lines corresponding to a plurality of electrical appliances 4, and sets the target period for identifying the power consumption pattern to a period prior to the present time.

[0041] Returning to FIG. 14, the notification information storage unit 2136 stores message information for generating notification information prompting each electrical appliance 4 to start or stop operation, message information for generating notification information prompting linkage settings for multiple electrical appliances 4 that have similar power consumption patterns, and message information for generating notification information notifying when the electrical appliance 4 will start or stop operating.

[0042] The operation period storage unit 2138 stores operation period information indicating the operation period of each of the plurality of electric devices 4 estimated based on the element vector in association with device identification information that identifies each electric device 4. Here, the device identification information includes, for example, an IP address or a MAC address of each of the plurality of electric devices 4.

[0043] The operation period estimation unit 2118 identifies an element vector with the largest weighting coefficient for each of the identified at least one power consumption pattern, and, based on the identified element vector, identifies an operation period to be estimated for the electric appliance 4. Then, the operation period estimation unit 2118 stores operation period information indicating the estimated operation period in the operation period storage unit 2138 in association with device identification information that identifies the electric appliance 4.

[0044] The notification time identification unit 2119 generates notification information prompting each of the plurality of electrical appliances 4 to start or stop operation and identifies the notification time to transmit the notification information to the terminal devices 51, 52 based on the operation time information of each of the plurality of electrical appliances 4 stored in the operation time storage unit 2138. The notification time identification unit 2119 identifies, for example, a time point a preset delay time after the operation start time or stop time indicated by the operation time information stored in the operation time storage unit 2138 as the notification time. Here, the delay time is set to, for example, 5 minutes. The notification time identification unit 2119 notifies the notification unit 2117 of the identified notification time.

[0045] The similar operation period determination unit 2120 determines whether there are multiple electric appliances 4 with similar operation periods. Specifically, the similar operation period determination unit 2120 calculates the errors in the start and end periods of the operation periods of each of the multiple electric appliances 4 stored in the operation period storage unit 2138. If both the errors in the calculated start and end periods of the operation periods are equal to or less than a preset error threshold, the similar operation period determination unit 2120 determines that the operation periods of the corresponding multiple electric appliances 4 are similar to each other. Furthermore, if the similar operation period determination unit 2120 determines that there are multiple operation periods that are similar to each other, it notifies the notification unit 2117 of the appliance identification information of the multiple electric appliances 4 with similar operation periods.

[0046] The control information generation unit 2121 generates control information for controlling the operation of the electric appliances 4 based on the operation periods indicated by the operation period information of each of the electric appliances 4 stored in the operation period storage unit 2138. Then, the control information generation unit 2121 notifies the control information transmission unit 2122 of the generated control information together with the device identification information of the electric appliance 4 corresponding to the control information stored in the operation period storage unit 2138. The control information transmission unit 2122 transmits the control information generated by the control information generation unit 2121 to the electric appliance 4 identified by the device identification information notified from the control information generation unit 2121.

[0047] Each time the notification time identified by the notification time identification unit 2119 arrives, the device state determination unit 2123 acquires the most recent power amount vector information stored in the vector storage unit 132 and element vector information indicating the element vector identified by the operation time estimation unit 2118, stored in the element information storage unit 134. Then, the device state determination unit 2123 determines whether the differential power amount between the power consumption equivalent to the element of the power amount vector corresponding to the notification time indicated by the power amount vector information and the power amount equivalent to the element of the element vector corresponding to the notification time is greater than a preset differential power amount threshold.

[0048] When the appliance state determination unit 2123 determines that the differential power amount is greater than the differential power amount threshold at the notification time, the notification unit 2117 transmits notification information to the terminal device 52 that prompts the terminal device 52 to start or stop operation of the electrical appliance 4. Furthermore, when the similar operation period determination unit 121 notifies the terminal device 52 of the appliance identification information of the plurality of electrical appliances 4, the notification unit 117 transmits notification information to the terminal device 52 that prompts the terminal device 52 to link the plurality of electrical appliances 4 identified by the notified appliance identification information. Furthermore, the notification unit 117 transmits notification information to the terminal device 52 that includes operation start time information indicating the operation start time of the electrical appliance 4 or stop time information indicating the stop time of the electrical appliance 4, based on the operation time information stored in the operation time storage unit 127.

[0049] Next, the information providing process executed by the cloud server 2001 according to this embodiment will be described with reference to Figures 16 and 17. Note that in Figures 16 and 17, the same processes as those in embodiment 1 are denoted by the same reference numerals as those in Figures 10 and 11. Furthermore, similar to embodiment 1, in parallel with the information providing process, the power amount obtaining unit 111 obtains the power amount information transmitted from the control device 2 by transmitting power amount information request information to the control device 2 requesting the control device 2 to transmit power amount information each time the power consumption information obtaining time arrives, and stores the obtained power amount information in the power amount storage unit 131.

[0050] First, as shown in FIG. 16, a series of processes from step S101 to S113 is executed. Next, the pattern number determination unit 115 determines whether the number Npa of the multiple power consumption patterns is equal to or greater than a preset pattern number threshold Npath (step S114). If the pattern number determination unit 115 determines that the number Npa of the multiple power consumption patterns is equal to or greater than the pattern number threshold Npath (step S114: Yes), the process of step S2108, which will be described later, is executed. On the other hand, suppose that the pattern number determination unit 115 determines that the number Npa of the multiple power consumption patterns is less than the pattern number threshold Npath (step S114: No). In this case, the operation time estimation unit 2118 identifies the power consumption pattern with the highest probability of occurrence in the most recent target period stored in the pattern storage unit 135 (step S2101).

[0051] Next, the operation period estimation unit 2118 identifies an element vector stored in the pattern storage unit 135 that has the largest average weighting coefficient indicated by the average weighting coefficient information corresponding to the identified power consumption pattern (step S2102). Based on the identified element vector, the operation period estimation unit 2118 then identifies the start and end of a period during which the power consumption P is equal to or greater than a preset power consumption threshold Pth, and stores operation period information indicating the operation period of the electric appliance 4 defined by the identified start and end in the operation period storage unit 2138 (step S2103). The notification unit 2117 then generates notification information notifying the terminal devices 51 and 52 of the operation start or stop period of the electric appliance 4, and transmits the generated notification information to the terminal devices 51 and 52 (step S2104). This notification information includes operation start period information indicating the operation start period of the electric appliance 4 or stop period information indicating the stop period of the electric appliance 4.

[0052] Next, the similar operation period determination unit 2120 calculates the error in each start date and end date of the operation period of each of the plurality of electric appliances 4 stored in the operation period storage unit 2138 (step S2105). Subsequently, the similar operation period determination unit 2120 determines whether both the error in the start date and the error in the end date of the calculated operation period are equal to or less than a preset error threshold (step S2106). Here, if the similar operation period determination unit 2120 determines that both the error in the start date and the error in the end date of the operation period are equal to or less than the error threshold (step S2106: No), the processing of step S2108 described below is executed. On the other hand, if the similar operation period determination unit 2120 determines that both the error in the start date and the error in the end date of the operation period are equal to or less than the error threshold (step S2106: Yes), it notifies the notification unit 2117 of the device identification information of the plurality of electric appliances 4 having similar operation periods. Then, the notification unit 117 generates notification information prompting the multiple electrical appliances 4 identified by the appliance identification information notified by the similar operating period determination unit 121 to be linked, and transmits the notification information to the terminal devices 51 and 52 (step S2107).

[0053] 17, the notification time identification unit 2119 and the control information generation unit 2121 refer to the operation time information of each of the plurality of electric appliances 4 stored in the operation time storage unit 2138, and determine whether the operation start time or stop time has arrived for any of the plurality of electric appliances 4 (step S2108). If the notification time identification unit 2119 and the control information generation unit 2121 determine that the operation start time or stop time has not yet arrived for all of the plurality of electric appliances 4 (step S2108: No), the process of step S101 is executed again. On the other hand, it is assumed that the notification time identification unit 2119 and the control information generation unit 2121 determine that the operation start time or stop time has arrived for any of the plurality of electric appliances 4 (step S2108: Yes). In this case, the notification time determination unit 2119 and the control information generation unit 2121 determine whether the electrical appliance 4 whose operation start time or stop time has arrived is set to an automatic control mode in which the start or stop control of operation is performed based on control information transmitted from the cloud server 2001 (step S2109).

[0054] Here, when the notification time identification unit 2119 determines that the electric appliance 4 is not set to the automatic control mode (step S2109: No), it identifies a notification time for transmitting notification information to the terminal device 52, prompting the user to start or stop the operation of the electric appliance 4, based on the operation time information stored in the operation time storage unit 2138 (step S2110). Thereafter, the appliance state determination unit 2123 determines whether the identified notification time has arrived (step S2111). Here, the appliance state determination unit 2123 repeatedly executes the process of step S2111 unless the notification time has arrived (step S2111: No). On the other hand, when the appliance state determination unit 2123 determines that the notification time has arrived (step S2111: Yes), it acquires the most recent power energy vector information stored in the vector storage unit 132 and element vector information indicating the element vector identified by the operation time estimation unit 2118, stored in the element information storage unit 134. Then, the appliance state determination unit 2123 determines whether the differential energy |P-Pc| between the power consumption P corresponding to the element of the power amount vector corresponding to the notification time indicated by the power amount vector information and the power amount Pc corresponding to the element of the element vector corresponding to the notification time is greater than a preset differential energy threshold Pmth (step S2112). If the appliance state determination unit 2123 determines that the differential energy |P-Pc| is equal to or less than the differential energy threshold Pmth (step S2112: No), the process of step S101 is executed again. On the other hand, it is assumed that the appliance state determination unit 2123 determines that the differential energy |P-Pc| is greater than the differential energy threshold Pmth (step S2112: Yes). In this case, the notification unit 2117 generates notification information prompting the terminal device 52 to start or stop operation of the electrical appliance 4 and transmits the notification information to the terminal device 52 (step S2113). Next, the process of step S101 is executed again.

[0055] Furthermore, in step S2109, if the control information generator 2121 determines that the electric appliance 4 is set to the automatic control mode (step S2109: Yes), the control information generator 2121 generates control information for starting or stopping the operation of the electric appliance 4. Then, the control information transmitter 2122 transmits the generated control information to the electric appliance 4 (step S2114). Subsequently, the process of step S101 is executed again.

[0056] As described above, in the cloud server 1 according to this embodiment, the operation time estimation unit 2118 identifies an element vector with a maximum weighting coefficient for the power consumption pattern identified by the pattern identification unit 114, and estimates the operation time of the electric appliance 4 based on the identified element vector. Then, the control information generation unit 2121 generates control information for controlling the operation of the electric appliance 4 based on the estimated operation time. This makes it possible to start or stop the operation of the electric appliance 4 in accordance with the behavior pattern of the user even when the user is not present in the home where the electric appliance 4 is installed, thereby improving the crime prevention effect of the home when the user is not present in the home where the electric appliance 4 is installed.

[0057] Furthermore, in the cloud server 2001 according to this embodiment, the notification time identification unit 2119 identifies a notification time for transmitting notification information to the terminal devices 51 and 52, prompting the user to start or stop operation of each of the electrical appliances 4, based on the operation time information. When the notification time arrives, the appliance state determination unit 2123 determines whether the differential power between the amount of power consumption corresponding to the element of the power vector corresponding to the notification time and the amount of power corresponding to the element of the element vector corresponding to the notification time is greater than a predetermined differential power threshold. If the notification unit 2117 determines that the differential power is greater than the differential power threshold, it transmits notification information to the terminal devices 51 and 52, prompting the user to start or stop operation of the electrical appliance 4. This allows, for example, a notification to the user that the user forgot to stop the electrical appliance 4 when the electrical appliance 4 is not operating during an operation period when the electrical appliance 4 is operating according to the user's normal lifestyle pattern, thereby preventing the electrical appliance 4 from operating unnecessarily. Furthermore, by transmitting notification information to the terminal device 51 owned by the user's administrator, the administrator can be notified promptly when an abnormality occurs in the user.

[0058] Furthermore, in the cloud server 2001 according to this embodiment, the similar operation period determination unit 2120 determines whether or not there are multiple operation periods that are similar to each other, and if it is determined that there are multiple operation periods that are similar to each other, the notification unit 2117 transmits notification information to the terminal devices 51 and 52 to prompt the terminal devices 51 and 52 to link the multiple electric appliances 4 that correspond to the multiple operation periods that are similar to each other. This makes it possible to provide the user with the optimal operation start and stop times for each of the multiple electric appliances 4.

[0059] Furthermore, the notification unit 2117 according to this embodiment transmits notification information including operation start time information indicating the operation start time or stop time information indicating the stop time of the electric appliance 4 to the terminal devices 51, 52, based on the operation time of the electric appliance 4 estimated by the operation time estimation unit 2118. This makes it possible to present to the user the optimal operation start time and stop time for each electric appliance 4. This makes it easier for the user to decide whether to reduce the amount of electricity purchased by shifting the operation time of the electric appliance 4 to a time period when a midnight electricity rate applies or a time period when surplus electricity is generated.

[0060] Although the embodiments of the present disclosure have been described above, the present disclosure is not limited to the above-described embodiments. For example, in the first embodiment, the notification determination unit 116 may determine whether the magnitude of a difference vector between an energy vector generated based on a weighting coefficient vector corresponding to a power consumption pattern to which a certain target period belongs and a plurality of element vectors and an energy vector indicating the actual energy consumption of the electrical appliance 4 during the target period is equal to or greater than a predetermined difference threshold. Here, the weighting coefficient vector is a weighting coefficient vector corresponding to at least one energy consumption pattern identified based on a plurality of energy vectors corresponding to each of a plurality of target periods prior to the certain target period. Then, when the notification unit 117 determines that the magnitude of the difference vector is equal to or greater than the difference threshold, the notification unit 117 may transmit notification information to the terminal devices 51 and 52 to notify the user that their lifestyle patterns have changed.

[0061] In the first embodiment, the vector generation unit 112 may generate power vectors for each of a plurality of target periods based on the weather information, each associated with the weather information. The hardware configuration of the cloud server according to this modification is the same as that described in the first embodiment. In the description of this modification, the same hardware components as those in the first embodiment will be described using the same reference numerals as those shown in FIG. 2 as appropriate. Also, as shown in FIG. 18, the cloud server 3001 according to this modification can communicate with a weather server 3008 via a wide area network NW1. The weather server 3008 manages weather information for each of a plurality of predefined regions in association with regional information indicating each of the regions. Here, the weather information includes average temperature information indicating the average daily temperature in each region and maximum temperature information indicating the maximum daily temperature. When the weather server 3008 receives weather information request information from the cloud server 3001, which requests the weather server 3008 to transmit weather information, the weather server 3008 identifies the weather information according to the received weather information request information. Here, the weather information request information includes area information indicating the area where the house where the electrical appliance 4 is installed is located, and the weather server 3008 identifies the weather information corresponding to the area information included in the weather information request information. Then, the weather server 3008 transmits the identified weather information to the cloud server 3001.

[0062] 18, the same components as those in the first embodiment are denoted by the same reference numerals as in FIG. 3. The auxiliary storage unit 103 includes an energy storage unit 131, a vector storage unit 3132, a processing target storage unit 3133, an element information storage unit 3134, a pattern storage unit 3135, a notification information storage unit 136, a terminal identification information storage unit 137, and a weather information storage unit 3139.

[0063] 19(A), for example, the weather information storage unit 3139 stores average temperature information indicating average temperatures over multiple periods and maximum temperature information indicating maximum temperatures, in association with period information indicating the corresponding periods and house identification information identifying the houses in which the electrical appliances 4 are installed. As shown in Fig. 19(B), for example, the processing target storage unit 3133 stores target period information defined based on weather conditions, branch line information identifying the branch lines L1, L2, L3, L4, and L5 for which power consumption amounts are to be obtained, and combinations of house identification information identifying the houses in which the branch lines L1, L2, L3, L4, and L5 are installed, in association with service identification information, element number information, and processing identification information.

[0064] The vector storage unit 3132 stores power energy vector information indicating a power energy vector for each target period, in association with average temperature information and maximum temperature information for that target period. The element information storage unit 3134 stores element vector information indicating multiple element vectors for each temperature range indicating a temperature range for the target period, and weight coefficient information indicating a weight coefficient corresponding to each element vector, in association with the processing identification information, temperature range information indicating the corresponding temperature range, and element identification information identifying each element vector, as shown in Fig. 20 for example. The pattern storage unit 3135 stores combinations of candidate pattern identification information identifying candidate power consumption patterns to which the element vector for each target period belongs, average weight coefficient information, and probability information indicating the probability that the element vector belongs to each candidate pattern, in association with the target period information and processing identification information, as shown in Fig. 21 for example.

[0065] 18 , the notification information storage unit 3136 stores multiple types of message information included in the notification information transmitted to the terminal devices 51 and 52. Here, the notification information storage unit 3136 stores the message information included in the notification information related to the settings of the electric appliance 4 in association with information indicating the range of the average power consumption of the electric appliance 4 in the target period.

[0066] The weather information acquisition unit 3124 acquires the weather information transmitted from the weather server 3008 by transmitting the above-mentioned weather information request information to the weather server 3008. Then, the weather information acquisition unit 3124 extracts the above-mentioned average temperature information and maximum temperature information from the acquired weather information, and stores the extracted average temperature information and maximum temperature information in the weather information storage unit 3139 in association with the season information and the house identification information.

[0067] The vector generation unit 3112 generates an electric energy vector corresponding to each of the aforementioned target periods using the electric energy information stored in the electric energy storage unit 131. Then, the vector generation unit 3112 stores the electric energy vector information indicating the generated electric energy vector in the vector storage unit 3132 in association with the average temperature information and maximum temperature information stored in the weather information storage unit 3139.

[0068] The element decomposition unit 3113 calculates, for each of a plurality of target periods, a plurality of weighting coefficients corresponding to each of the plurality of element vectors obtained by decomposing a plurality of power energy vectors into the plurality of element vectors. Next, the element decomposition unit 3113 classifies element vector information indicating the calculated plurality of element vectors and weighting coefficient information indicating the corresponding plurality of weighting coefficients into a plurality of preset temperature ranges based on the average temperature information and maximum temperature information for each of the plurality of target periods. Next, the element decomposition unit 3113 calculates the average value of the weighting coefficients indicated by the weighting coefficient information for each temperature range. Then, the element decomposition unit 3113 stores the element vector information belonging to each temperature range and the average weighting coefficient information indicating the corresponding average weighting coefficient in the element information storage unit 3134 in association with the element identification information, temperature range information indicating the temperature range, and processing identification information.

[0069] The pattern identification unit 3114 identifies the power consumption pattern for each of the multiple target periods by clustering the multiple target periods into multiple power consumption patterns using a weighting coefficient vector whose elements are multiple weighting coefficients corresponding to each of the multiple target periods.The pattern identification unit 114 then stores average weighting coefficient vector information indicating the average weighting coefficient vector of the power consumption pattern for each target period and corresponding probability information indicating the corresponding probability in the pattern storage unit 3135 in association with the pattern identification information, target period information, temperature range information indicating the temperature range to which the average temperature or maximum temperature for the target period belongs, and processing identification information.

[0070] When the notification determination unit 116 determines that the power consumption pattern at the most recent first time point has transitioned to a different power consumption pattern from a time point earlier than the first time point, the notification unit 3117 identifies the power consumption pattern with the highest probability of occurrence within the most recent target period stored in the pattern storage unit 135. Here, the notification determination unit 116 may identify the power consumption pattern for each target period during a target period between the first time point and a third time point that is a predetermined number of target periods prior to the most recent first time point, and determine the power consumption pattern with the highest occurrence frequency within the target period as the power consumption pattern at the first time point. This allows the notification determination unit 116 to avoid determining that a transition to a different power consumption pattern has occurred when, for example, a sudden change in the power consumption pattern occurs within only one target period, thereby more appropriately determining whether or not a power consumption pattern has transitioned. Furthermore, the notification unit 3117 identifies the element vector with the highest average weighting coefficient indicated by the average weighting coefficient information corresponding to the identified power consumption pattern stored in the pattern storage unit 135, and calculates the average power consumption based on the identified element vector. Then, the notification unit 3117 selects message information corresponding to the calculated power consumption from the message information stored in the notification information storage unit 3136, and uses the selected message information to generate notification information regarding a setting change of the electrical appliance 4 and transmits it to the terminal devices 51 and 52.

[0071] Next, an information providing process executed by the cloud server 3001 according to this modification will be described with reference to FIG. 22. Note that in FIG. 22, the same processes as those in the first embodiment are denoted by the same reference numerals as those in FIGS. 10 and 11. First, a series of processes from steps S101 to S116 is executed. Next, the notification determination unit 116 refers to the pattern identification information stored in the pattern storage unit 135 and determines whether the power consumption pattern at the most recent first time point has transitioned to a power consumption pattern different from the power consumption pattern prior to the first time point (step S117). Here, the notification determination unit 116 may identify the power consumption pattern of each target period in the determination target period between the first time point and a third time point that is a preset number of target periods prior to the most recent first time point, as described above, and determine whether the power consumption pattern has transitioned when the power consumption pattern with the highest frequency of occurrence within the determination target period is the power consumption pattern at the first time point. Then, when the notification determination unit 116 determines that the power consumption pattern before the first time point is maintained constant (step S117: No), the process of step S101 is executed again.

[0072] On the other hand, suppose that the notification determination unit 116 determines that the power consumption pattern at the most recent first time point has transitioned to a different power consumption pattern from a time point earlier than the first time point (step S117: Yes). In this case, the notification unit 3117 identifies the power consumption pattern with the highest probability of occurrence in the most recent target period stored in the pattern storage unit 135 (step S3101). Next, the notification unit 3117 identifies the element vector with the highest average weighting coefficient indicated by the average weighting coefficient information corresponding to the identified power consumption pattern stored in the pattern storage unit 135 (step S3102). Next, the notification unit 3117 calculates the average power consumption based on the identified element vector (step S3103). Thereafter, the notification unit 3117 selects message information corresponding to the calculated power consumption from the message information stored in the notification information storage unit 3136, and uses the selected message information to generate notification information related to a setting change of the electrical appliance 4 and transmits the generated notification information to the terminal devices 51 and 52 (step S3104). Next, the process of step S101 is executed again.

[0073] According to this configuration, notification information can be generated taking into consideration the weather conditions in the area where the house is located during each target period.

[0074] In the first embodiment, an example has been described in which the pattern identification unit 114 identifies the power consumption pattern for each of a plurality of target periods by clustering into a plurality of power consumption patterns a weight coefficient vector whose elements are a plurality of weight coefficients corresponding to each of the target periods calculated by the element decomposition unit 113. However, the present invention is not limited to this, and for example, the pattern identification unit may identify the power consumption pattern for each of a plurality of target periods by clustering the power amount vector generated by the vector generation unit 112 into a plurality of power consumption patterns.

[0075] The hardware configuration of the cloud server according to this modification is the same as the hardware configuration described in the first embodiment. In the description of this modification, the same hardware configuration as in the first embodiment will be described using the same reference numerals as those shown in FIG. 2 as appropriate. As shown in FIG. 23, the CPU 101 of the cloud server 4001 according to this modification reads a program stored in the auxiliary storage unit 103 into the main storage unit 102 and executes it, thereby functioning as an energy amount acquisition unit 111, a vector generation unit 112, a pattern identification unit 4114, a number of patterns determination unit 115, a notification determination unit 116, a notification unit 117, and an average vector calculation unit 4125. Note that in FIG. 21, the same components as in the first embodiment are denoted by the same reference numerals as in FIG. 3. The auxiliary storage unit 103 also includes an energy amount storage unit 131, a vector storage unit 132, a processing target storage unit 133, a pattern storage unit 135, a notification information storage unit 136, a terminal identification information storage unit 137, and an average vector storage unit 4139. The average vector storage unit 4139 stores average power vector information indicating the average power vector of the power vectors belonging to each of a plurality of power consumption patterns in association with power consumption pattern identification information that identifies the corresponding power consumption pattern.

[0076] The pattern identification unit 4114 identifies the power consumption pattern for each of the multiple target periods by clustering the power amount vector for each of the multiple target periods into multiple power consumption patterns. Furthermore, when the pattern identification unit 4114 has identified the multiple power consumption patterns, it calculates the average power consumption for each of the multiple power consumption patterns in each of the multiple time periods. Then, if the average, median, or maximum value of the difference in the average power consumption for each of the multiple time periods between two power consumption patterns selected from the multiple power consumption patterns is equal to or less than a predetermined difference threshold, the pattern identification unit 4114 executes pattern integration processing to integrate the two selected power consumption patterns into one power consumption pattern.

[0077] The average vector calculation unit 4125 calculates, for each of the plurality of power consumption patterns after the pattern integration process, an average power energy vector of the power energy vectors belonging to each power consumption pattern, by referring to the power energy vectors stored in the vector storage unit 132. Then, the average vector calculation unit 4125 stores average power energy vector information indicating the calculated average power energy vector in the average vector storage unit 4139 in association with power consumption pattern identification information that identifies the corresponding power consumption pattern.

[0078] Next, an information providing process executed by the cloud server 4001 according to this modification will be described with reference to FIG. 24. Note that in FIG. 24, the same processes as those in the first embodiment are denoted by the same reference numerals as those in FIGS. 10 and 11. First, the vector generation unit 112 determines whether or not a pattern confirmation time has arrived for confirming a preset power consumption pattern (step S101). Here, if the vector generation unit 112 determines that the pattern confirmation time has arrived (step S101: Yes), it identifies power energy information to be used from the power energy information stored in the power energy storage unit 131 (step S102). Next, the pattern identification unit 4114 clusters the generated power energy vectors corresponding to the multiple target periods into multiple power consumption patterns (step S4101). As a result, the pattern identification unit 4114 identifies the power consumption pattern for each of the multiple target periods. Next, the pattern identification unit 114 calculates an average power energy vector of the power energy vectors for a target period belonging to each of the plurality of power consumption patterns, and calculates a difference vector between the calculated plurality of average power energy vectors (step S110). Then, the pattern identification unit 114 determines whether or not there are a plurality of average power energy vectors whose magnitude dP of the difference vector is equal to or less than a predetermined difference threshold dPth (step S111). If the pattern identification unit 114 determines that there is no average power energy vector whose magnitude |dVp| of the difference vector is equal to or less than the difference threshold |dVpth| (step S111: No), the process of step S4102, which will be described later, is executed. On the other hand, if the pattern identification unit 114 determines that there are a plurality of average power energy vectors whose magnitude dP of the difference vector is equal to or less than a predetermined difference threshold dPth (step S111: Yes), the pattern identification unit 114 executes an integration process to integrate the power consumption patterns corresponding to the plurality of average power energy vectors into a single power consumption pattern (step S112). Thereafter, the average vector calculation unit 4125 calculates the average power amount vector of the power amount vectors belonging to each power consumption pattern for each of the plurality of power consumption patterns after the pattern integration process described above.Then, the average vector calculation unit 4125 stores average power vector information indicating the calculated average power vector in the average vector storage unit 4139 in association with power consumption pattern identification information that identifies the corresponding power consumption pattern (step S4102). After that, a series of processes from step S114 onwards are executed.

[0079] According to this configuration, the process of decomposing the power amount vector into a plurality of element vectors is not executed, and therefore the processing load on the cloud server 4001 can be reduced accordingly.

[0080] In each embodiment, an example has been described in which the element decomposition unit 113 calculates the average, median, or maximum value of the error of the power amount vector for all target periods included in the target period group, and adopts the number of element vectors and the corresponding weighting coefficients such that the calculated average, median, or maximum value of the error is less than a predetermined error threshold. However, this is not limiting, and the element decomposition unit 113 may, for example, determine the number of element vectors using the Akaike Information Criterion (AIC) or the Bayesian Information Criterion (BIC), and adopt the determined number of element vectors and the corresponding weighting coefficients.

[0081] In each embodiment, an example has been described in which the length of the target period is one day (24 hours), but the length of the target period is not particularly limited, and may be, for example, two days, one week, or the like.

[0082] Furthermore, the various functions of cloud servers 1, 2001, 3001, and 4001 according to the present disclosure may be realized by software, firmware, or a combination of software and firmware. In this case, the software or firmware may be written as a program, and the program may be stored and distributed on a computer-readable recording medium such as a flexible disk, a CD-ROM (Compact Disc Read Only Memory), a DVD (Digital Versatile Disc), or an MO (Magneto-Optical Disc). The program may then be read and installed on a computer to configure a computer capable of realizing each of the aforementioned functions. Furthermore, if each function is realized by sharing the work between an operating system (OS) and an application, or by cooperation between the OS and the application, only the parts other than the OS may be stored on the recording medium.

[0083] Furthermore, each program can be superimposed on a carrier wave and distributed over a network. For example, the program can be posted on a bulletin board system (BBS) on the network and distributed over the network. These programs can then be started and run under the control of the OS in the same way as other application programs, thereby enabling the above-mentioned processing to be performed. [Industrial Applicability]

[0084] The present disclosure is suitable for a service that provides information to a user of an electrical device according to the user's behavioral patterns. [Explanation of symbols]

[0085] 1,2001,3001,4001 Cloud server, 2 Control device, 4 Electrical equipment, 6 BBR, 21 Distribution board, 22 Power meter, 51,52 Terminal device, 101 CPU, 102 Main memory unit, 103 Auxiliary memory unit, 106 Wide area communication unit, 108 Timekeeping unit, 109 Bus, 111 Power amount acquisition unit, 112,3112 Vector generation unit, 113,3113 Element decomposition unit, 114,3114,4114 Pattern identification unit, 115 Pattern number determination unit, 116 Notification determination unit, 117,2117,3117 Notification unit, 120 Operation time identification unit, 131 Power amount memory unit, 132,3132 Vector memory unit, 133,3133 Processing target memory unit, 134,3134 Element information storage unit, 135, 3135 Pattern storage unit, 136, 2136, 3136 Notification information storage unit, 137 Terminal identification information storage unit, 2118 Operation time estimation unit, 2119 Notification time specification unit, 2120 Similar operation time determination unit, 2121 Control information generation unit, 2122 Control information transmission unit, 2123 Equipment status determination unit, 2138 Operation time storage unit, 3124 Weather information acquisition unit, 3139 Weather information storage unit, 4125 Mean vector calculation unit, 4139 Mean vector storage unit, NW1 Wide area network, NW2 Local network, PS System power supply

Claims

1. a vector generation unit that generates a plurality of power amount vectors, each of which corresponds to a plurality of target periods for which at least one power consumption pattern is to be identified, and each of which has as elements an amount of power consumption consumed by at least one electrical device in each of a plurality of time periods that make up the target periods; an element decomposition unit that calculates, for each of the plurality of target periods, a plurality of weighting coefficients corresponding to each of the plurality of element vectors obtained by decomposing the plurality of power amount vectors into a plurality of element vectors; a pattern identification unit that identifies the power consumption pattern for each of the plurality of target periods by clustering the plurality of target periods into a plurality of power consumption patterns using a weight coefficient vector having the plurality of weight coefficients corresponding to the plurality of target periods as elements; a notification unit that generates notification information for notifying a user of content related to the power consumption pattern based on the power consumption pattern for each of the plurality of target periods and transmits the notification information to a terminal device. Information provision device.

2. When the pattern identification unit has identified a plurality of the power consumption patterns, it calculates an average power consumption amount for each of the plurality of power consumption patterns in each of the plurality of time periods, and when an average, median, or maximum value of a difference in the average power consumption amounts for each of the plurality of time periods between two of the power consumption patterns selected from the plurality of power consumption patterns is equal to or less than a predetermined difference threshold, it integrates the two selected power consumption patterns into one power consumption pattern. The information providing device according to claim 1 .

3. the vector generation unit generates a plurality of power amount vectors whose elements are amounts of power consumed by a plurality of electrical devices in each of a plurality of time periods constituting the target period; 3. The information providing device according to claim 1.

4. a notification determination unit that determines whether or not at least one first power consumption pattern identified based on a plurality of power amount vectors corresponding to a first time point and a plurality of target periods in the past before the first time point differs from at least one second power consumption pattern identified based on a plurality of power amount vectors corresponding to a plurality of target periods in the past before a second time point that is a preset set period in the past before the first time point; When it is determined that the first power consumption pattern and the second power consumption pattern are different, the notification unit transmits the notification information to the terminal device to notify the user that a lifestyle pattern of the user has changed. The information providing device according to any one of claims 1 to 3.

5. a notification determination unit that determines whether or not a magnitude of a difference vector between an electric energy vector generated based on the weight coefficient vector corresponding to at least one of the electric energy patterns identified based on a plurality of electric energy vectors corresponding to each of a plurality of target periods prior to a first time point and the plurality of element vectors, and the electric energy vector at the first time point, is equal to or greater than a predetermined difference threshold; When the notification unit determines that the magnitude of the difference vector is equal to or greater than the difference threshold, the notification unit transmits the notification information to the terminal device to notify that a change has occurred in the user's lifestyle pattern. The information providing device according to any one of claims 1 to 3.

6. a terminal identification information storage unit configured to store terminal identification information for identifying at least one terminal device to which the notification information is to be sent; the notification unit transmits the notification information to at least one terminal device identified by the terminal identification information stored in the terminal identification information storage unit; The information providing device according to any one of claims 1 to 5.

7. an operation timing estimation unit that identifies the element vector with the largest weighting coefficient for each of the at least one identified power consumption pattern, and estimates an operation timing for the at least one electrical device based on the identified element vector; a control information generation unit that generates control information for controlling an operation of the at least one electrical device based on the estimated operation time. The information providing device according to any one of claims 1 to 6.

8. an operation timing estimation unit that identifies the element vector with the largest weighting coefficient for each of the at least one identified power consumption pattern, and estimates an operation timing for the at least one electrical device based on the identified element vector; a notification time specifying unit that specifies a notification time for transmitting notification information prompting the terminal device to start or stop operation of each of the at least one electric appliance based on the operation time; and a device state determination unit that, when the notification time arrives, determines whether or not a differential power amount between the power consumption amount corresponding to the element of the power amount vector corresponding to the notification time and the power amount corresponding to the element of the element vector corresponding to the notification time is greater than a predetermined differential power amount threshold, When it is determined that the power difference amount is greater than the power difference amount threshold, the notification unit transmits the notification information to the terminal device, prompting the terminal device to start or stop operation of the at least one electric appliance. The information providing device according to any one of claims 1 to 6.

9. the pattern identification unit identifies at least one of the power consumption patterns for each of a plurality of electrical devices; an operation timing estimation unit that identifies the element vector with the largest weighting coefficient for at least one of the power consumption patterns identified for each of the plurality of electrical devices, and that determines a predicted operation timing for each of the plurality of electrical devices based on the identified element vector; a similar operation period determination unit that determines whether or not there are multiple operation periods that are similar to each other, When it is determined that there are a plurality of operation periods similar to each other, the notification unit transmits the notification information to the terminal device, prompting the terminal device to interlock a plurality of electrical appliances corresponding to the plurality of operation periods similar to each other. The information providing device according to any one of claims 1 to 8.

10. The notification unit transmits, to the terminal device, operation start time information indicating an operation start time or stop time information indicating a stop time of the at least one electrical appliance based on the operation time.

10. The information providing device according to claim 8 or 9.

11. a vector generation unit that generates a plurality of power amount vectors, each of which corresponds to a plurality of target periods for which at least one power consumption pattern is to be identified, and each of which has as elements an amount of power consumption consumed by at least one electrical device in each of a plurality of time periods that make up the target periods; a pattern identification unit that identifies the power consumption pattern for each of the plurality of target periods by clustering the plurality of target periods into a plurality of power consumption patterns using the plurality of power amount vectors, and when the plurality of power consumption patterns have been identified, calculates an average power consumption amount for each of the plurality of power consumption patterns in each of the plurality of time periods, and, when an average, median or maximum value of a difference in the average power consumption amounts for each of the plurality of time periods between two of the power consumption patterns selected from the plurality of power consumption patterns is equal to or less than a predetermined difference threshold, integrates the two selected power consumption patterns into one power consumption pattern; a notification determination unit that determines whether to notify a user of notification information related to the power consumption pattern based on the power consumption pattern for each of the plurality of target periods; a notification unit that, when it is determined that the notification information is to be notified, identifies the notification information to be notified. Information provision device.

12. a pattern number determination unit that determines whether or not a number of the plurality of power consumption patterns obtained by clustering the plurality of target periods into the plurality of power consumption patterns is equal to or greater than a predetermined pattern number threshold; the notification unit, when determining that the number of patterns is equal to or greater than the pattern number threshold, avoids transmitting the notification information to the terminal device. The information providing device according to any one of claims 1 to 11.

13. generating a plurality of power amount vectors, each of which corresponds to a plurality of target periods for which at least one power consumption pattern is to be identified, and each of which has as its elements the amount of power consumed by at least one electrical device in each of a plurality of time periods constituting the target periods; calculating, for each of the plurality of target periods, a plurality of weighting coefficients corresponding to each of the plurality of element vectors obtained by decomposing the plurality of power amount vectors into a plurality of element vectors; identifying the power consumption pattern for each of the plurality of target periods by clustering the plurality of target periods into a plurality of power consumption patterns using a weight coefficient vector having the plurality of weight coefficients corresponding to the plurality of target periods as elements; generating notification information for notifying a user of content related to the power consumption pattern based on the power consumption pattern for each of the plurality of target periods, and transmitting the notification information to a terminal device. Information provision method.

14. Computer, a vector generation unit that generates a plurality of power amount vectors, each of which corresponds to a plurality of target periods for which at least one power consumption pattern is to be identified, and each of which has as elements an amount of power consumption consumed by at least one electrical device in each of a plurality of time periods that make up the target periods; an element decomposition unit that calculates, for each of the plurality of target periods, a plurality of weighting coefficients corresponding to each of the plurality of element vectors obtained by decomposing the plurality of power amount vectors into a plurality of element vectors; a pattern identification unit that identifies the power consumption pattern for each of the plurality of target periods by clustering the plurality of target periods into a plurality of power consumption patterns using a weight coefficient vector having the plurality of weight coefficients corresponding to the plurality of target periods as elements; a notification unit that generates notification information for notifying a user of content related to the power consumption pattern based on the power consumption pattern for each of the plurality of target periods, and transmits the notification information to a terminal device; A program to function as a

Images