Programs, information processing methods, information processing devices

Abstract

To provide an information processing device that can obtain more accurate animal weight data. [Solution] The information processing device 30 comprises a measurement information acquisition unit 331 and a data extraction unit 332. The measurement information acquisition unit 331 performs an acquisition process to acquire weight measurement data from a measuring device 20 that measures the weight of a pet. The data extraction unit 332 performs an extraction process to extract a group of pet weight data by excluding outliers from the weight measurement data group that should not be extracted as pet weight data.

Description

Technical Field

[0001] The present invention relates to a program, an information processing method, and an information processing apparatus.

Background Art

[0002] Conventionally, there is an information processing system described in Patent Document 1 below. This information processing system includes a measuring device and an information processing device. The measuring device is incorporated in a pet toilet used by a pet and measures the weight of the pet. The information processing device estimates the pet's medical condition based on the weight measured using the measuring device.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In an information processing system such as that described in Patent Document 1, since animals such as pets often do not get on the measuring device in a state suitable for weight measurement, there is a possibility that the measurement data of the weight includes values that deviate significantly from the true value. Also, due to inappropriate installation conditions of the measuring device, there is a possibility that the measuring device measures abnormal values. When using a data group of measurement data including outliers such as values that deviate significantly from the true value and abnormal values, there is a possibility that the weight value of the pet cannot be appropriately detected.

[0005] The present invention has been made in view of such circumstances, and an object thereof is to provide a program, an information processing method, and an information processing apparatus capable of obtaining more appropriate weight data of an animal.

Means for Solving the Problems

[0006] The program that solves the above problem causes the computer to perform an acquisition process to obtain weight measurement data from a measuring device that measures the weight of animals, and an extraction process to extract a data set of animal weight data by removing outliers that should not be extracted as animal weight data from the data set of measurement data.

[0007] The information processing method that solves the above problem involves a computer performing an acquisition process to obtain weight measurement data from a measuring device that measures the weight of animals, and an extraction process to extract a data set of animal weight data by removing outliers from the data set of measurement data that should not be extracted as animal weight data.

[0008] An information processing device that solves the above problems comprises an acquisition unit that acquires weight measurement data from a measuring device that measures animals, and an extraction unit that extracts a data set of animal weight data by removing outliers from the data set of measurement data that should not be extracted as animal weight data.

[0009] These configurations make it possible to obtain more accurate weight data. [Effects of the Invention]

[0010] According to the program, information processing method, and information processing device of the present invention, it is possible to obtain more appropriate animal weight data. [Brief explanation of the drawing]

[0011] [Figure 1] A block diagram showing the schematic configuration of the information processing system of the embodiment. [Figure 2] A block diagram showing the schematic configuration of the information processing device of the embodiment. [Figure 3] (A) and (B) are graphs showing examples of measurement data sets for the embodiment. [Figure 4] A diagram showing an example of a pattern in the measurement data set of the embodiment. [Figure 5] (A) and (B) are graphs showing examples of measurement data sets for the embodiment. [Figure 6] (A) and (B) are graphs showing examples of measurement data groups of the embodiment. [Figure 7] (A) and (B) are graphs showing examples of measurement data groups of the embodiment. [Figure 8] A flowchart showing the procedure of the learning process of the learning model of the embodiment. [Figure 9] A flowchart showing the procedure of the process executed by the control unit of the embodiment. [Figure 10] A diagram showing an example of a screen displayed on the user terminal of the embodiment. [Figure 11] (A) and (B) are diagrams showing examples of images displayed on the user terminal of the embodiment. [Figure 12] A diagram showing an example of a screen displayed on the user terminal of the embodiment. [Figure 13] (A) and (B) are diagrams showing examples of images displayed on the user terminal of the embodiment. [Figure 14] A sequence chart showing an operation example of the information processing system of the embodiment. [Figure 15] A block diagram showing the hardware configuration of the computer of the embodiment. [Figure 16] A diagram showing an example of an image displayed on the user terminal of the first modification of the embodiment. [Figure 17] A flowchart showing the procedure of the process executed by the control unit of the second modification of the embodiment. [Figure 18] A flowchart showing the procedure of the process executed by the control unit of the second modification of the embodiment. [Figure 19] A flowchart showing the procedure of the process executed by the control unit of the third modification of the embodiment.

Mode for Carrying Out the Invention

[0012] Hereinafter, an embodiment of a program, an information processing method, and an information processing apparatus will be described with reference to the drawings. To facilitate understanding of the description, the same reference numerals are given to the same components in each drawing as much as possible, and duplicate descriptions are omitted.

[0013] <First Embodiment> First, an overview of the information processing system of this embodiment will be described.

[0014] (Overview of Information Processing System) As shown in FIG. 1, the information processing system 10 of this embodiment includes a measurement device 20, an information processing device 30, and a user terminal 40. The measurement device 20, the information processing device 30, and the user terminal 40 are communicably connected to each other via a communication network N10.

[0015] The measuring device 20 is incorporated into a pet toilet 50 used by pets. Pets are animals such as cats and dogs. The pet toilet 50 is a portable container made of synthetic resin, for example, with an opening for pets to enter and exit. The measuring device 20 is equipped with various sensors and can acquire various data related to pets. These sensors include, for example, a weight sensor capable of measuring the weight of an object placed on the pet toilet 50. When the measuring device 20 detects a change in weight, for example, using the weight sensor, it determines that a pet has entered the toilet 50. After detecting that a pet has entered the toilet, if that state continues for a predetermined time, the measuring device 20 measures the weight using the weight sensor. The predetermined time is, for example, 30 seconds. Alternatively, the measuring device 20 may measure the weight at the time it detects that a pet has entered the toilet 50. Typically, the measuring device 20 transmits measurement information to the information processing device 30 via the communication network N10 each time it measures weight, but is not limited to this. It may also store the measurement information as batch processing for a predetermined period (e.g., daily or weekly) and transmit the stored measurement information to the information processing device 30. The measurement information includes measurement data of the weight measured by the measuring device 20, the date and time of measurement, the user ID, and information about the equipment used. The user ID is identification information assigned individually to each user. The information about the equipment used is information that can identify the type of pet toilet used by the user, such as the model number of the pet toilet. The measurement information may also include a pet ID. The pet ID is identification information assigned individually to each pet.

[0016] The information processing device 30 is, for example, a server computer, a cloud computer, or a personal computer. The information processing device 30 stores the weight measurement data included in the measurement information transmitted from the measuring device 20 in a database. The information processing device 30 extracts a data set of pet weights from the data set of weight measurement data stored in the database by excluding outliers that should not be extracted as pet weight data. Outliers are, for example, values ​​that are far from the true value of the pet's weight. Therefore, the weight data set consists of multiple weight measurement data that are estimated to be close to the true value of the pet's weight. Hereinafter, the data set of weight measurement data will be referred to as the measurement data set, and the data set of pet weights will be referred to as the weight data set.

[0017] The information processing device 30 further performs processing using the extracted weight data set. This processing may include, for example, estimating the pet's health status based on the weight data set, or estimating the pet's weight by analyzing the weight data set. After performing processing using the weight data set, the information processing device 30 transmits the data processing results to the user terminal 40 via the communication network N10. For example, if processing to estimate the pet's health status is performed as part of the processing using the weight data set, the data processing result will be the estimated result of the pet's health status. Also, for example, if analysis processing of the weight data set is performed as part of the processing using the weight data set, the data processing result will be the estimated value of the pet's weight obtained through the analysis processing.

[0018] The user terminal 40 is, for example, a smartphone, a tablet, or a personal computer. The user terminal 40 is a device owned by a user, for example, a pet owner. The user terminal 40 has a web browser and a dedicated application that corresponds to the system 10 installed. The user can perform various operations on the system 10 by launching the dedicated application or by accessing a dedicated website that corresponds to the system 10 via a web browser. Various operations on the system 10 include, for example, inputting various information about the user and various setting operations related to processing using weight data sets. The user can also view various information provided by the information processing device 30 via the dedicated application or dedicated website. For example, when the user terminal 40 receives data processing results transmitted from the information processing device 30, it displays the received data processing results via the dedicated application or dedicated website. Specifically, the user terminal 40 displays, as data processing results, estimated results of the pet's health status and estimated values ​​of the pet's weight. By viewing the estimated health status and estimated weight of the pet displayed on the user terminal 40, the user can more easily determine whether or not the pet is healthy, thereby improving user convenience. In this embodiment, the user terminal 40 corresponds to the user's terminal device.

[0019] (Configuration of information processing device) Next, the configuration of the information processing device 30 will be described.

[0020] As shown in Figure 2, the information processing device 30 comprises a communication unit 31, a storage unit 32, and a control unit 33.

[0021] The communication unit 31 transmits and receives various information to and from the measuring device 20 and the user terminal 40 via the communication network N10.

[0022] The storage unit 32 stores various types of information held by the information processing device 30. For example, the storage unit 32 stores various programs executed by the control unit 33. The storage unit 32 also stores the user database 320, the measurement database 321, and the learning model 323.

[0023] The user database 320 stores user information for each of the multiple users of this system 10, including user ID, password, equipment information, declared number of pets, pet ID, pet name, and declared pet weight. The user ID and password are transmitted from the user terminal 40, for example, when the user accesses a dedicated application or website, and are used for user authentication in the information processing device 30. The declared number of pets is the total number of pets owned by each user, as declared by the user. The declared pet weight is the weight of each pet owned by each user, as declared by the user. If there are multiple declared pets, the declared pet weight includes the weight information for each of the multiple pets. As an example of storing information about each pet separately from user information, each pet may be stored with associated pet information such as pet ID, pet name, pet breed, pet age, declared pet weight, and statistically processed values ​​(e.g., estimated weight) from measurement data sets and / or weight data sets. Pet information may be associated with user information, for example, by pet ID.

[0024] The measurement database 321 stores weight measurement data, measurement date and time, and body weight data associated with each user ID and equipment information of multiple users. The weight measurement data is the weight data measured by the measuring device 20. The body weight data is data obtained by excluding outliers that are far from the true value of the body weight from the measurement data; in other words, it is data that is estimated to be close to the true value of the pet's body weight. In this embodiment, multiple measurement data stored in the measurement database 321 are used as a measurement data group, and multiple body weight data stored in the measurement database 321 are used as a body weight data group.

[0025] The learning model 323 is a model constructed using machine learning techniques (including deep learning; the same applies hereinafter). The learning model 323 is constructed, for example, by a neural network. The learning model 323 performs, for example, the process of extracting weight data from measurement data and the process of identifying patterns (hereinafter simply referred to as "patterns") in the measurement data or weight data, and outputs the results of these processes. The process of extracting weight data takes measurement data stored in the measurement DB 321 for a predetermined period Ta as input data and removes outliers from the measurement data. The predetermined period Ta is, for example, two months.

[0026] The patterns may include, for example, a normal pattern indicating that the measurement data set or weight data set is normal, and an error pattern indicating that there is an error in the measurement data set or weight data set. The patterns may also include, as other examples, an increase pattern indicating a trend of weight gain in a pet, and / or a decrease pattern indicating a trend of weight loss in a pet. Furthermore, the increase patterns may include a normal increase pattern indicating a normal increase trend and an abnormal increase pattern indicating an abnormal increase trend. Similarly, the decrease patterns may include a normal decrease pattern indicating a normal decrease trend and an abnormal decrease pattern indicating an abnormal decrease trend.

[0027] The input data (explanatory variables) to be input to the learning model 323 could be, for example, a set of measurement data or each measurement data, and pet information (e.g., pet breed) and / or information about the equipment used that is associated with the set of measurement data or each measurement data.

[0028] The control unit 33 shown in Figure 2 controls the information processing device 30. The control unit 33 has a functional configuration realized by executing a program stored in the storage unit 32, and includes a registration information acquisition unit 330, a measurement information acquisition unit 331, a data extraction unit 332, a determination unit 333, a data processing execution unit 334, and a notification unit 335.

[0029] When input data such as a set of measurement data for Ta over a predetermined period is received by the learning model 323, it may perform at least one of the following classifications (1) to (3). Then, the data extraction unit 332 may extract a set of weight data from the set of measurement data based on the result of this classification. (1) Classify whether a group of weight data located within a specific area is included in the group of measurement data. (2) Classify whether each data point in the measurement data group is weight data or not. (3) Classify which of the multiple patterns the measurement data group belongs to.

[0030] Instead of the classification in (1) above, for example, the data set may be classified based on whether it was measured under conditions suitable for weight measurement, or whether it is necessary or sufficient to perform subsequent data processing. Also, the difference between the classifications in (1) and (2) above is that (1) classifies by the data set, while (2) classifies each data item that makes up the data set individually.

[0031] The data (target variable) output by the learning model 323 may be information indicating the result of at least one of the classifications (1) to (3) above. For example, in the case of classification (3) above, the information indicating the classification result may be the probability for each pattern. In this case, it is assumed that the data was classified into the pattern with the highest probability.

[0032] The training data for the learning model 323 may include a dataset of the above input data and ground truth data which is information representing at least one of the above classifications (1) to (3) corresponding to the above output data. For example, in the case of (1) above, the input data may be a group of measurement data, and the ground truth data may be information indicating whether the group of measurement data includes weight data.

[0033] The learning model 323 analyzes the measurement data set and, if there is a region where multiple measurement data are arranged side by side along the time axis, it determines (classifies) that a trend line has been formed and that the weight data set located within that specific region is included in the measurement data set.

[0034] For example, suppose a set of measurement data shown in Figure 3(A) is input to the learning model 323 as a set of measurement data for a predetermined period Ta. In the set of measurement data shown in Figure 3(A), multiple measurement data are arranged in the time axis direction within region A10. In such a case, the learning model 323 determines that the measurement data within region A10 is weight data, while the measurement data in regions other than region A10 is determined to be outliers. The data extraction unit 332 then extracts the measurement data within region A10 as weight data. As a result, the data extraction unit 332 uses the learning model 323 to extract one trend line, or in other words, one set of weight data. Note that in the graphs in Figures 3, 5 to 7, the vertical axis is weight (in grams) and the horizontal axis is the measurement date and time (in date).

[0035] The data extraction unit 332 may determine, for example, that it has extracted a (necessary or sufficient) set of weight data if the extracted weight data set satisfies predetermined conditions (also referred to as "weight data conditions"). The weight data conditions may include, for example, that the number of data points in the weight data set is greater than a set threshold, and / or that the ratio of the weight data set to the measurement data set is greater than a set threshold.

[0036] Furthermore, let's assume that the measurement data set shown in Figure 3(B) is input to the learning model 323 as the measurement data set for a predetermined period Ta. The measurement data set shown in Figure 3(B) contains two regions A20 and A21 in which multiple measurement data are arranged side by side in the time axis direction. Therefore, the learning model 323 determines that each measurement data within regions A20 and A21 forms a trend line for regions A20 and A21, and thus determines that they are weight data sets. Consequently, the learning model 323 determines that two weight data sets can be extracted. In this case, the learning model 323 determines that the measurement data in regions other than regions A20 and A21 are outliers. The data extraction unit 332 then extracts the data set from which the outliers have been removed, i.e., the measurement data sets within regions A20 and A21, as weight data sets. As a result, the data extraction unit 332 uses the learning model 323 to extract two trend lines, or in other words, two weight data sets.

[0037] On the other hand, when the data extraction unit 332 extracts one or more weight data sets using the learning model 323 as shown in Figures 3(A) and (B), it compares the number of extracted weight data sets with the number of pets declared in the user DB 320. If they match, it determines that it has been possible to extract the weight data sets.

[0038] For example, the data extraction unit 332 determines that it has successfully extracted a weight data set if it can extract one weight data set from the measurement data set shown in Figure 3(A) using the learning model 323, and the number of pets reported by the user is one. Furthermore, the learning model 323 determines that it has successfully extracted two weight data sets if it can extract two weight data sets from the measurement data set shown in Figure 3(B), and the number of pets reported by the user is two.

[0039] Furthermore, even if the data extraction unit 332 was able to extract two weight data sets from the measurement data set shown in Figure 3(B), it still determines that it was able to extract a weight data set if the declared number of pets is one, i.e., if the number of extracted weight data sets is greater than the declared number of pets. In this case, although there is an error in the number of pets declared by the user, the two extracted weight data sets are considered to be correct, and therefore the data extraction unit 332 determines that it was able to extract a weight data set. In such a situation, of the two weight data sets, the one that is closest to the declared weight of the pet stored in the user DB 320 is extracted as the weight data set. For example, when two weight data sets were able to extract from the measurement data set shown in Figure 3(B), if the declared weight of the pet is 3500g, the measurement data set in area A21 is extracted as the weight data set.

[0040] Areas A10, A20, and A21 could be, for example, (a) areas set based on the declared weight received from the user's operation (e.g., within ±500g of the declared weight), (b) areas received from the user's operation (e.g., areas specified via the screen), or (c) one or more areas identified based on weight data of animals such as pets that have been measured and accumulated over a predetermined period using learning techniques such as machine learning.

[0041] In addition to or instead of the classification (1) to (3) above by the learning model 323, the data extraction unit 332 may extract weight data sets based on whether the measurement data sets satisfy predetermined conditions (also called "extractable conditions"). As extractable conditions, for example, any of the following conditions (a1) to (a3), or an AND condition thereof, can be used.

[0042] (a1) The variability of the measurement data set is less than or equal to a predetermined value. Specifically, this variability is the range of weight values ​​that the measurement data set of Ta can take over a predetermined period (the difference between the maximum and minimum values). The predetermined value is, for example, 200g. Alternatively, instead of the condition that the variability of the measurement data set is less than or equal to a predetermined value, the condition that the variance or standard deviation of the measurement data set within a specific area is less than or equal to a predetermined value may be used.

[0043] (a2) The number of declared pets stored in the user DB320 is two or more, and the difference in declared weights of each pet is greater than or equal to a predetermined value. The predetermined value is, for example, 500g.

[0044] (a3) Measurement data exists prior to the most recent specified period Ta, and measurement data exists for a total of more than a specified percentage of the total number of days, and there are no consecutive periods within the specified period during which measurement data does not exist. The specified percentage is, for example, 50%, and the specified period is, for example, two weeks.

[0045] As described above, if a weight data set can be extracted from a set of measurement data for a predetermined period of Ta, that measurement data set can be classified into one of the normal patterns P10 or P11, for example, as shown in Figure 4. The ideal pattern shown in Figure 4 is, for example, the case where there are the same number of trend lines as the number of declared pets. If the learning model 323 is able to extract a weight data set from a set of measurement data for a predetermined period of Ta, it determines whether the extracted weight data set corresponds to one of the normal patterns P10 or P11. The learning model 323 then outputs a determination result indicating whether the measurement data set corresponds to one of the normal patterns P10 or P11, either together with or individually to the extracted weight data set.

[0046] On the other hand, depending on the content of the data set measured by the measuring device 20 in Ta for a predetermined period, it may not be possible to extract the necessary or sufficient weight data set from the measurement data set to perform subsequent data processing. Figures 5 to 7 show examples of measurement data sets from which weight data set cannot be extracted in this manner.

[0047] If it is not possible to extract weight data from the measurement data set of Ta for a predetermined period, the measurement data set can be classified into one of the error patterns P20 to P26 shown in Figure 4, for example. When the data extraction unit 332 determines that it is not possible to extract weight data from the measurement data set of Ta for a predetermined period, it determines which of the error patterns P20 to P26 the measurement data set corresponds to and outputs the determination result.

[0048] Figure 5(A) shows an example of a measurement data set when there is an overall data shortage. When the data extraction unit 332 (including the learning model 323) acquires a measurement data set like the one shown in Figure 5(A) as a measurement data set for a predetermined period Ta, it determines that it cannot extract the weight data set, for example, based on the fact that the number of measurement data included in the measurement data set for the predetermined period Ta is less than or equal to a predetermined number. In such a case, the data extraction unit 332 may determine that the measurement data set corresponds to error pattern P20.

[0049] Figure 5(B) shows an example of a measurement data set where data is partially missing. In Figure 5, the area where data is missing is enclosed by a dashed line A30. When the data extraction unit 332 acquires a measurement data set like the one shown in Figure 5(B) as a measurement data set for a predetermined period Ta, it determines that it cannot extract the weight data set, for example, based on the fact that there is no measurement data for a certain period within the measurement data set for the predetermined period Ta. In such a case, the data extraction unit 332 may determine that the measurement data set corresponds to error pattern P21.

[0050] Figure 6(A) shows an example of a set of measurement data when a user stops using the system 10. Such a set of measurement data contains areas where no data exists, as indicated by the area enclosed by the dashed line A31 in Figure 6(A). When the data extraction unit 332 acquires a set of measurement data like the one shown in Figure 6(A) as a set of measurement data for a predetermined period Ta, it determines that it cannot extract the weight data set, for example, based on the fact that no measurement data exists for a part of the predetermined period Ta. This part of the period may be, for example, the latter half of the predetermined period Ta (a part indicating a certain period from the end of the period). In such a case, the data extraction unit 332 may determine that the set of measurement data corresponds to error pattern P22.

[0051] Figure 6(B) shows an example of a measurement data set when there is a partial weight variation exceeding a predetermined value. When the data extraction unit 332 acquires a measurement data set like the one shown in Figure 6(B) as the measurement data set for a predetermined period Ta, it calculates the variance or standard deviation based on the measurement data set. If the calculated result does not satisfy the predetermined conditions, the data extraction unit 332 determines that there is a variation in the measurement data set for the predetermined period Ta and that it cannot extract the weight data set. Instead of calculating the variance or standard deviation, the data extraction unit 332 determines whether there is a variation exceeding a predetermined value. If the data extraction unit 332 determines that there is a variation exceeding a predetermined value, it determines that it cannot extract the weight data set. In such cases, the data extraction unit 332 may determine that the measurement data set corresponds to error pattern P23.

[0052] Figure 7(A) shows an example of a measurement data set when there is overall weight variation. When the data extraction unit 332 acquires a measurement data set like the one shown in Figure 7(A) as a measurement data set for a predetermined period Ta, it determines that it cannot extract the weight data set based, for example, on the absence of a trend line in the measurement data set for the predetermined period Ta, or on the presence of a fluctuation exceeding a predetermined value. The predetermined value is, for example, 500g. In such cases, the data extraction unit 332 may determine that the measurement data set corresponds to error pattern P24.

[0053] Figure 7(B) shows an example of a measurement data set when an abnormal increase or decrease occurs. When the data extraction unit 332 acquires a measurement data set like the one shown in Figure 7(B) as a measurement data set for a predetermined period Ta, it determines that it cannot extract the weight data set, for example, based on the fact that a difference greater than a predetermined value occurs in the measurement data set. The predetermined value is set to a value in the range of |500g| to |2000g|. In such a case, the data extraction unit 332 may determine that the measurement data set corresponds to error pattern P25.

[0054] Although not shown in the diagram, the data extraction unit 332 determines that it cannot extract a weight data group even if it was able to extract one weight data group from the measurement data group shown in Figure 3(A), for example, if the number of pets reported by the user is two, that is, if the number of pets reported by the user is greater than the number of weight data groups extracted.

[0055] In addition, the information processing device 30 may have multiple learning models 323, each corresponding to a different type of pet toilet 50 (specifically, a different model number, etc.), stored in the memory unit 32. In this case, the same learning model 323 would be used for the same type of pet toilet. Alternatively, the information processing device 30 may have multiple learning models 323, each corresponding to a different user, stored in the memory unit 32.

[0056] Next, the learning procedure for the learning model 323 will be described. Figure 8 is a flowchart showing an example of the learning procedure for the learning model 323. Note that the process shown in Figure 8 may be executed at a predetermined interval, for example, one week.

[0057] As shown in Figure 8, during the training of the learning model 323, a dataset used for training the learning model 323 is first generated (step S10). The dataset is generated by annotation, which involves assigning ground truth data called labels to each of the multiple measurement data. Ground truth data is information indicating whether each of the multiple measurement data is weight data or not. As multiple measurement data to which ground truth data is assigned, for example, measurement data corresponding to each of the multiple patterns P10, P11, P20~26 shown in Figure 4 are used.

[0058] Next, the dataset generated in step S10 is input to the learning model 323 as training data, thereby training the learning model 323, for example, using deep learning (step S11). Then, the trained learning model 323 is registered in the memory unit 32 (step S12).

[0059] The registration information acquisition unit 330 acquires user registration information from the user terminal 40 via the communication unit 31. User registration information includes user ID, password, equipment information, declared number of pets, and declared weight of pets. The registration information acquisition unit 330 stores the registration information acquired from the user terminal 40 in the user database 320, associating it with the user ID.

[0060] The measurement information acquisition unit 331 acquires measurement information transmitted from the measuring device 20 via the communication unit 31. The measurement information includes measurement data of weight measured by the measuring device 20, measurement date and time, user ID, and equipment information used. Each time the measurement information acquisition unit 331 acquires measurement information from the measuring device 20, it stores the weight measurement data included in the acquired measurement information in the measurement DB 321, associating it with the user ID and equipment information used. The measurement information may include a pet ID instead of a user ID.

[0061] The data extraction unit 332, the determination unit 333, the data processing execution unit 334, and the notification unit 335 perform various processes on the measurement data stored in the measurement DB 321. Figure 9 is a flowchart showing an example of the procedure for the processes performed by the data extraction unit 332, the determination unit 333, the data processing execution unit 334, and the notification unit 335. The processes shown in Figure 9 are performed at predetermined intervals, which are, for example, one week (weekly) or one day (daily). The data extraction unit 332 may, for example, have the learning model 323 perform some or all of the processes.

[0062] As shown in Figure 9, the data extraction unit 332 first acquires a group of measurement data for Ta over a predetermined period from the measurement DB 321 (step S20). Then, the data extraction unit 332 inputs the acquired group of measurement data for Ta over a predetermined period as input data into the learning model 323 to extract a group of weight data and perform pattern discrimination (step S21).

[0063] At this time, if the data extraction unit 332 is able to extract a group of weight data from the group of measurement data of Ta over a predetermined period using the learning model 323, it determines which of the normal patterns P10 and P11 shown in Figure 4 the extracted weight data group corresponds to, and outputs the extracted weight data group along with the determination result.

[0064] On the other hand, if the data extraction unit 332 cannot extract the weight data set from the measurement data set of Ta for a predetermined period using the learning model 323, it determines which of the error patterns P20 to P26 the measurement data set corresponds to and outputs the determination result.

[0065] Next, the determination unit 333 determines whether the discrimination result of the learning model 323 indicates a normal pattern (step S22). If the determination unit 333 determines that the discrimination result of the learning model 323 indicates a normal pattern (step S22: YES), that is, if it was possible to extract the weight data group from the measurement data group of Ta for a predetermined period, the data processing execution unit 334 performs processing using the extracted weight data group (step S23). Processing using the weight data group may include, for example, estimating the health status of the pet or calculating the estimated weight of the pet from the weight data group. The estimated weight is obtained by performing predetermined statistical processing on the weight data group, for example, by calculating statistical values ​​of the weight data group (e.g., mean, median, etc.).

[0066] The data processing execution unit 334, as part of the process of estimating the pet's health status, determines that the pet's health status has deteriorated if, for example, the estimated weight of the pet in the most recent week has decreased by a predetermined percentage or increased by a predetermined percentage compared to the estimated weight of the pet in the previous month. The predetermined percentage is, for example, 20%. The data processing execution unit 334 may also, as part of the process of estimating the pet's health status, if it determines that the pet's weight is in an increasing or decreasing pattern, estimate that the weight is on an increasing or decreasing trend based on the determined pattern.

[0067] Next, the notification unit 335 notifies the user terminal 40 of the processing results of the data processing execution unit 334, i.e., the estimated health status and / or estimated weight of the pet (step S24). As one form of notification, the notification unit 335 may generate display information including, for example, the health status and / or estimated weight, and transmit it to the user terminal 40. The display information is information for the user terminal 40 to display various information in a web browser or a dedicated application corresponding to this system 10. The notification unit 335 may also notify the user terminal 40 of expert comments corresponding to the pet's health status. Furthermore, the notification unit 335 may transmit display information to the user terminal 40 including the measurement data set and / or weight data set used to estimate the pet's health status. This display information may include, for example, a plot image in which the data set is plotted, as shown in Figures 3,5 to 7. The plot image may be displayed in a manner that allows for the identification of the weight data set and outliers. Specifically, the color and shape of the weight data plot may be different from the color and shape of the outlier plot. This allows the user terminal 40 to display statistical values ​​and plot images of the measurement data set and weight data set used for estimating the pet's health status, etc. For example, the user can check from the plot state of the data set whether the measuring device 20 is working properly, whether the pet is on the measuring device 20 in a suitable position for weight measurement, etc. In this embodiment, one example of notification processing is when the notification unit 335 notifies the user terminal 40 of the processing result of the data processing execution unit 334, and one example of output processing is when the user terminal 40 outputs the processing result.

[0068] In step S22, if the determination unit 333 determines that the learning model 323's discrimination result does not show a normal pattern (step S22: NO), that is, if the measurement data group shows an error pattern, the notification unit 335 sends an error notification to the user terminal 40 (step S25). The error notification is information to be displayed on the user terminal 40, such as an error message indicating that the weight data group cannot be acquired, or advice on how to resolve the error pattern in the discrimination result. In this embodiment, the process by which the notification unit 335 sends an error notification to the user terminal 40 is an example of a notification process, and the process of outputting that error notification from the user terminal 40 is an example of an output process.

[0069] Next, the method by which the notification unit 335 notifies the user terminal 40 will be explained. For the following explanation, we will use the case where the user's pet is a cat as an example.

[0070] Figure 10 shows an example of a normal notification screen 70 that is displayed on the user terminal 40 when the information processing device 30 notifies the user terminal 40 of the pet's health status during the process of step S24 shown in Figure 9.

[0071] As shown in Figure 10, this normal notification screen 70 is provided with a pet name display area 701, a measurement period display area 702, a comment display area 703, an estimated status display area 704, and a weight data display area 705.

[0072] The pet's name is displayed in the pet name display area 701. The measurement period display area 702 displays the period during which the measurement data set was measured (or the period during which the measurement data set was acquired). The comment display area 703 displays expert comments corresponding to the pet's health condition.

[0073] The estimated status display area 704 displays the pet's health status estimated by the data processing execution unit 334. If the pet's health status is estimated to be normal, the estimated status display area 704 displays a message indicating that the pet's health is normal, as shown in Figure 10. On the other hand, if the pet's health status is estimated to be normal, the estimated status display area 704 displays, for example, the estimated health status of the pet, the estimated illness of the pet, weight gain or loss, and a message encouraging the pet to seek medical attention, as shown in Figures 11(A) and (B).

[0074] As shown in Figure 10, the estimated weight of the pet calculated by the data processing execution unit 334 is displayed in the weight data display area 705. The average weight for the previous month and the average weight for this week may also be displayed in the weight data display area 705 in a way that allows for comparison, such as using a bar graph.

[0075] Figure 12 shows an example of the error notification screen 71 displayed on the user terminal 40 when an error notification is sent from the information processing device 30 to the user terminal 40 during the process of step S25 shown in Figure 9.

[0076] As shown in Figure 12, the error notification screen 71, like the normal notification screen 70 shown in Figure 10, is provided with a pet name display area 711, a measurement period display area 712, a comment display area 713, an estimated status display area 714, and a weight data display area 715.

[0077] The pet name display area 711 and the measurement period display area 712 display the pet name and measurement period, respectively, just like on the normal notification screen 70. On the other hand, the comment display area 713 displays a message indicating that no comment will be displayed if weight data cannot be obtained. Also, unlike the normal notification screen 70, the weight data display area 715 does not display estimated weight data, etc.

[0078] The estimated state display area 714 displays a message corresponding to the error pattern identified by the learning model 323 from among multiple error patterns P20 to P26. For example, if the error pattern identified by the learning model 323 is P20 shown in Figure 4, the estimated state display area 714 displays a message as shown in Figure 12. In contrast, if the error pattern identified by the learning model 323 is either P21 or P22 shown in Figure 4, the estimated state display area 714 displays a message as shown in Figure 13(A). Furthermore, if the error pattern identified by the learning model 323 is P24 shown in Figure 4, the estimated state display area 714 displays a message as shown in Figure 13(B).

[0079] (Example of information processing system operation) Next, an example of the operation of the information processing system 10 of this embodiment will be described.

[0080] As shown in Figure 14, in the information processing system 10, first, the measuring device 20 transmits measurement information to the information processing device 30 each time it measures weight (step S30). When the information processing device 30 receives the measurement information transmitted from the measuring device 20 (step S40), it stores the weight measurement data and measurement date and time, etc., contained in the received measurement information in the measurement DB 321 (step S41). As a result, weight measurement data is accumulated in the measurement DB 321.

[0081] Next, when a predetermined period Ta has elapsed, the information processing device 30 reads the measurement data set for the predetermined period Ta from the measurement DB 321 (step S42), and inputs the read measurement data set for the predetermined period Ta into the learning model 323 (step S43). At this time, if the learning model 323 determines that the measurement data set shows a normal pattern, the information processing device 30 performs processing using the extracted weight data set (step S44), and notifies the user terminal 40 of the pet's health status and estimated weight (step S45). On the other hand, if the learning model 323 determines that the measurement data set shows an error pattern, the information processing device 30 sends an error notification to the user terminal 40 (step S46).

[0082] (Hardware configuration of the information processing system) Next, referring to Figure 15, an example of the hardware configuration of the computer 100 when the measuring device 20, information processing device 30, and user terminal 40 are implemented by the computer 100 will be described.

[0083] As shown in Figure 15, the computer 100 includes a processor 101, a storage device 102, a communication device 103, an input device 104, and an output device 105, etc. The processor 101 is a CPU (Central Processing Unit) or a GPU (Graphical Processing Unit), etc. The storage device 102 consists of at least one of, for example, memory, an HDD (Hard Disk Drive), and an SSD (Solid State Drive). The communication device 103 performs wired communication or wireless communication. The input device 104 is a device that accepts input operations and consists of at least one of, for example, a keyboard, a touch panel, a mouse, and a microphone. The output device 105 is a device that outputs information and consists of at least one of, for example, a display, a touch panel, and a speaker.

[0084] (Operation and effects of information processing systems) As described above, the information processing device 30 comprises a measurement information acquisition unit 331 and a data extraction unit 332. The measurement information acquisition unit 331 performs an acquisition process to acquire weight measurement data from a measuring device 20 that measures the biological information (weight) of a pet (animal). The data extraction unit 332 performs an extraction process to extract a group of pet weight data by excluding outliers from the weight measurement data group that should not be extracted as pet weight data. Outliers are, for example, values ​​that are far from the true value of the pet's weight.

[0085] This configuration makes it possible to obtain more accurate pet weight data.

[0086] The information processing device 30 further comprises a data processing execution unit 334. The data processing execution unit 334 performs processing using the pet weight data set based on the extraction result of the pet weight data set by the data extraction unit 332. This processing includes the result that normal patterns P10 and P11 were extracted, or the result that error patterns P20 to P26 were extracted. In other words, this processing includes the result that the pet weight data set was successfully extracted by the data extraction unit 332, or the result that the pet weight data could not be extracted by the data extraction unit 332.

[0087] This configuration makes it easy to determine whether or not the pet's weight data set has been successfully extracted.

[0088] The information processing device 30 further includes a determination unit 333. The determination unit 333 performs a determination process to determine whether the measurement data group is normal data or not, based on the extraction results of the data extraction unit 332. The information processing device 30 further includes a notification unit 335. If the determination unit 333 determines that the measurement data group has error patterns P20 to P26, in other words, if the measurement data group is determined not to be normal data, the notification unit 335 sends an error regarding the measurement data group to the user terminal 40. The notification unit 335 notifies the user terminal 40 of at least one of the following: an image of a graph in which the measurement data group is plotted, and information regarding the solution to the error.

[0089] With this configuration, if an error exists in the measurement data set, the user terminal 40 will notify the user of this fact, and an image of the graph of the measurement data set and a solution to the error will also be notified to the user terminal 40, thereby improving user convenience.

[0090] If the determination unit 333 determines that the measurement data set has normal patterns P10 and P11, in other words, if the measurement data set is determined to be normal data, the notification unit 335 notifies the user terminal 40 of the estimated health status of the pet. The notification unit 335 also notifies the user terminal 40 of expert comments (predetermined comments) based on the estimated health status of the pet. Furthermore, the notification unit 335 may notify the user terminal 40 of at least one of the pet's weight data set, the estimated weight (statistical value) calculated from the weight data set, and an image of a graph on which the measurement data set is plotted.

[0091] With this configuration, if the measurement data set is normal, the user terminal 40 will notify the user of the pet's health status, expert comments, estimated weight, and graph images, thereby improving user convenience.

[0092] The data extraction unit 332 extracts weight data from the measurement data set using the trained model 323. The trained model 323 is a model that has been trained using the measurement data set as input data and training data in which each data point in the measurement data set is information indicating whether or not it is the weight of a pet.

[0093] This configuration makes it possible to extract weight data sets more appropriately from measurement data sets.

[0094] (First variation) Next, a first modified example of the information processing system 10 of the above embodiment will be described.

[0095] The method for extracting weight data sets from measurement data sets is not limited to the method using the learning model 323 as in the above embodiment; for example, a method of extraction by user specification, specifically by user selection, may also be employed.

[0096] For example, in the modified information processing system 10, as shown by the dashed line in Figure 2, the control unit 33 further includes a reception unit 336. The reception unit 336 performs reception processing to receive user operations performed on the user terminal 40. Specifically, when an image of a graph plotting a group of measurement data for Ta over a predetermined period is displayed on the user terminal 40, the reception unit 336 receives an operation to select a group of pet weight data from that graph image. For example, when an image of a graph 72, as shown in Figure 16, is displayed on the user terminal 40, if the user performs an operation to enclose a part of the image 72 as shown by the dashed line L10 in Figure 16, the reception unit 336 receives a user operation to specify the group of measurement data contained within the area enclosed by the dashed line L10. In this case, the data extraction unit 332 extracts the group of measurement data present in the area enclosed by the dashed line L10 in the image 72 as a group of weight data.

[0097] Furthermore, the user may be able to specify multiple regions from an image of a graph plotting a set of measurement data for Ta over a predetermined period, that is, to select multiple sets of weight data.

[0098] As described above, in this modified information processing system 10, the notification unit 335 performs output processing to output information including the measurement data group to the user by sending and displaying the information including the measurement data group to the user terminal 40. The reception unit 336 performs reception processing to receive a specification from the user whether or not each data in the data group displayed on the user terminal 40 is pet weight data. The data extraction unit 332 extracts the weight data group based on the user specification received by the reception unit 336.

[0099] Even with this configuration, it is possible to extract weight data from measurement data.

[0100] (Second variation) Next, a second modified example of the information processing system 10 of the above embodiment will be described.

[0101] In this modified example, the control unit 33 executes the process shown in Figure 17 instead of the process shown in Figure 9. In the process shown in Figure 17, the same reference numerals are used for processes that are the same as those shown in Figure 9, so redundant explanations are omitted.

[0102] As shown in Figure 17, in the modified information processing system 10, following the processing in step S20, the determination unit 333 of the control unit 33 classifies the measurement data group of Ta for a predetermined period into normal patterns P10, P11 and error patterns P20 to P26 using conditional branching processing, without using the learning model 323 (step S50). For example, the processing shown in Figure 18 is used as the conditional branching processing.

[0103] As shown in Figure 18, in the conditional branching process, the determination unit 333 first determines whether the measurement data group of Ta for a predetermined period satisfies predetermined normal pattern conditions (step S60). As normal pattern conditions, for example, the AND conditions (a1) to (a3) ​​above are used. If the measurement data group of Ta for a predetermined period satisfies predetermined normal pattern conditions (step S60: YES), the determination unit 333 determines that the measurement data group of Ta for a predetermined period has normal patterns P10 and P11 shown in Figure 4 (step S61). Note that the processing order of the conditional branches shown in Figure 18 may be changed as appropriate, and multiple conditional branches may be made in parallel or consolidated, or at least one of the conditional branches may be omitted.

[0104] When the measurement data group for the predetermined period Ta does not satisfy the predetermined normal pattern condition (step S60: NO), the determination unit 333 determines whether there is no data only in the most recent predetermined period in the measurement data group for the predetermined period Ta (step S62). The predetermined period is, for example, one week. When there is no data only in the most recent predetermined period in the measurement data group for the predetermined period Ta (step S62: YES), the determination unit 333 determines that the measurement data group for the predetermined period Ta has the error patterns P21 and P22 shown in FIG. 4 (step S63).

[0105] When there is data in the most recent predetermined period in the measurement data group for the predetermined period Ta (step S62: NO), the determination unit 333 determines whether the total number of data in the measurement data group for the predetermined period Ta is less than or equal to a predetermined number (step S64). When the total number of data in the measurement data group for the predetermined period Ta is less than or equal to the predetermined number (step S64: YES), the determination unit 333 determines that the measurement data group for the predetermined period Ta has the error pattern P20 shown in FIG. 4 (step S65).

[0106] When the total number of data in the measurement data group for the predetermined period Ta exceeds the predetermined number (step S64: NO), the determination unit 333 determines whether the correlation coefficient r of the measurement data group for the predetermined period satisfies the relationship "A < r < B" with respect to the predetermined values A and B (step S66). Note that the predetermined values A and B have the relationship "A < B". When the correlation coefficient r satisfies the relationship "A < r < B" (step S66: YES), the determination unit 333 determines that the measurement data group for the predetermined period Ta has the error pattern P24 shown in FIG. 4 (step S67).

[0107] When the correlation coefficient r does not satisfy the relationship "A < r < B" (step S66: NO), the determination unit 333 determines whether the correlation coefficient r satisfies the relationship "C < r < D" with respect to the predetermined values C and D (step S68). Note that the predetermined values A, B, C, and D have the relationship "A < C < D < B". When the correlation coefficient r satisfies the relationship "C < r < D" (step S68: YES), the determination unit 333 determines that there is one trend line. In this case, the determination unit 333 determines whether the number of pets declared by the user is two (step S69). When the declared number of pets declared by the user is two (step S69: YES), it is determined that the measurement data group for the predetermined period Ta has the error pattern P26 shown in FIG. 4 (step S70). On the other hand, when the number of pets declared by the user is one (step S69: NO), the determination unit 333 determines that the measurement data group for the predetermined period Ta has the error pattern P25 shown in FIG. 4 (step S71).

[0108] Also, when the correlation coefficient r does not satisfy the relationship "C < r < D" (step S68: NO), the determination unit 333 determines that the measurement data group for the predetermined period Ta has the error pattern P23 shown in FIG. 4 (step S72).

[0109] As shown in Figure 17, following the processing in step S50, the determination unit 333 determines whether the measurement data group of Ta for a predetermined period has normal patterns P10 and P11 (step S51). If the determination process in step S61 is performed in the process shown in Figure 18, the determination unit 333 determines that the measurement data group of Ta for a predetermined period has normal patterns P10 and P11 (step S51: YES). In this case, the data extraction unit 332 extracts the weight data group by inputting the measurement data group of Ta for a predetermined period as input data to the learning model 323 (step S52). Subsequently, the processes in steps S23 and S24 are executed. That is, the data processing execution unit 334 performs processing using the extracted weight data group (step S23), and the notification unit 335 notifies the user terminal 40 of the processing results of the data processing execution unit 334, i.e., the estimated health status and estimated weight of the pet (step S24).

[0110] On the other hand, if any of the determination processes in steps S63, S65, S67, S70, S71, or S72 are executed in the process shown in Figure 18, the determination unit 333 determines that the measurement data group of Ta for a predetermined period has error patterns P20 to P26 (step S51: NO). In this case, the process in step S25 is executed. That is, the notification unit 335 sends an error notification to the user terminal 40 (step S25).

[0111] According to the configuration of the modified information processing system 10, it is only necessary to construct a learning model 323 that has the function of extracting weight data by excluding outliers from the measurement data, making it easier to construct the learning model 323.

[0112] (Third variation) Next, a third modified example of the information processing system 10 of the above embodiment will be described.

[0113] In this modified information processing system 10, when weight measurement data is acquired by the measuring device 20, the pet's health status is estimated in real time. The control unit 33 of the information processing device 30 in this modified version executes the process shown in Figure 19, for example.

[0114] As shown in Figure 19, in the control unit 33, when the measurement information acquisition unit 331 acquires measurement information from the measuring device 20 (step S80), the determination unit 333 determines whether the weight measurement data included in the acquired measurement information is body weight data or not (step S81). For example, the determination unit 333 inputs the weight measurement data into a predetermined learning model and obtains a determination result from the learning model of whether the weight measurement data is body weight data or not.

[0115] If the determination unit 333 determines that the weight measurement data is body weight data (step S82: YES), the data processing execution unit 334 performs processing using the body weight data (step S83). Processing using the body weight data set is, for example, a process to estimate the health status of the pet. Subsequently, the notification unit 335 notifies the user terminal 40 of the processing result of the data processing execution unit 334, for example, the estimated health status of the pet (step S84).

[0116] If the determination unit 333 determines that the weight measurement data is not body weight data (step S82: NO), the notification unit 335 sends an error notification to the user terminal 40 (step S85). The error notification includes, for example, an error message indicating that body weight data could not be obtained.

[0117] <Other Embodiments> This disclosure is not limited to the specific examples given above.

[0118] For example, the weight data set extracted from the measurement data set by the information processing device 30 is not limited to pet weight data, but can be any animal weight data set. The same applies to the weight data estimated by the information processing device 30.

[0119] The functional configuration of the information processing device 30 may be provided separately for the measuring device 20 only, the user terminal 40 only, or for the measuring device 20, the information processing device 30, and the user terminal 40.

[0120] Even the above-mentioned examples, with appropriate design modifications by those skilled in the art, are included within the scope of this disclosure, as long as they possess the features of this disclosure. The elements, their arrangement, conditions, shapes, etc., of each of the above-mentioned examples are not limited to those exemplified and can be modified as appropriate. The elements of each of the above-mentioned examples can be combined in different ways as appropriate, as long as no technical inconsistencies arise. [Explanation of Symbols]

[0121] 20: Measuring device, 30: Information processing device, 40: User terminal (terminal device), 100: Computer, 323: Learning model, 331: Measurement information acquisition unit, 332: Data extraction unit, 333: Judgment unit, 334: Data processing execution unit, 335: Notification unit, 336: Reception unit.

Claims

1. On the computer, A data acquisition process to obtain weight measurement data from a measuring device used to measure animals, The system performs an extraction process to extract the data set of animal weight data by removing outliers from the aforementioned measurement data set that should not be extracted as animal weight data. program.

2. The outliers are values ​​that are far from the true weight of the animal. The program according to claim 1.

3. To the aforementioned computer, Output processing that outputs the aforementioned measurement data set to the user, The process further involves receiving a request from the user to specify whether or not each data point in the data set output by the output process is data on the weight of the animal, The extraction process extracts the weight data set based on the user's specifications received by the reception process. The program according to claim 1.

4. To the aforementioned computer, Based on the extraction results of the weight data set by the extraction process, a process is executed using the weight data set. The program according to claim 1.

5. The result of extracting the weight data set includes either a result indicating that the weight data set could be extracted by the extraction process, or a result indicating that the weight data set could not be extracted by the extraction process. The program according to claim 4.

6. To the aforementioned computer, Based on the results of the extraction process, a determination process is further executed to determine whether the data set of measurement data is normal data. If the determination process determines that the data set of measurement data is not normal data, a notification process is further executed to notify the user's terminal device of the error related to the data set of measurement data. The program according to claim 1.

7. To the aforementioned computer, In the notification process, the user's terminal device is further notified of at least one of the following: an image of a graph plotting the data set of measurement data, and information regarding the solution to the error. The program according to claim 6.

8. To the aforementioned computer, As a process using the aforementioned weight data set, the system will perform a process to estimate the animal's health status based on the weight data set and notify the user's terminal device of the estimated health status of the animal. The program according to claim 1.

9. To the aforementioned computer, As a process using the aforementioned weight data set, the system will execute a process that further notifies the user's terminal device of a predetermined comment based on the estimated health status of the animal. The program according to claim 8.

10. To the aforementioned computer, As a process using the aforementioned weight data set, the user's terminal device is notified of at least one of the following: the weight data set, statistical values ​​based on the weight data set, and an image of a graph plotting the measurement data set. The program according to claim 1.

11. To the aforementioned computer, In the extraction process described above, a learning model is used to extract the weight data set from the measurement data set. The program according to claim 1.

12. The aforementioned learning model is a model trained using the data set of measurement data as input data, and training data in which each data point in the data set contains information indicating whether or not it represents the weight of an animal. The program according to claim 11.

13. Computers A data acquisition process to obtain weight measurement data from a measuring device used to measure animals, The following steps are performed: an extraction process is performed to extract the data set of animal weight data by removing outliers from the data set of measurement data that should not be extracted as animal weight data. Information processing methods.

14. An acquisition unit that acquires weight measurement data from a measuring device used to measure animals, The system includes an extraction unit that extracts a data set of animal weight data by removing outliers from the data set of measurement data that should not be extracted as animal weight data. Information processing device.

Images