Generation apparatus, generation method, and generation program

Abstract

To generate a model that more accurately predicts the onset of labor in pregnant women. [Solution] The generation device according to the present invention comprises an acquisition unit and a generation unit, wherein the acquisition unit acquires biometric data including heart rate or pulse rate from a user, and the generation unit uses as training data the biometric data features corresponding to each day from the date of onset of labor to the date of onset of labor, including the day after the date in which the biometric data features were at their lowest, and which are labeled as indicating whether or not the date after a predetermined number of days is the date of onset of labor, to generate a predictive model that predicts the onset of labor after the predetermined number of days.

Description

Technical Field

[0004] ,

[0006] , , , , , ,

[0005] , , ,

[0003] , , , , , ,

[0001] The present invention relates to a generation device, a generation method, and a generation program.

Background Art

[0002] Conventionally, a doctor uses the due date of a pregnant woman, an echocardiogram, etc. to estimate the 0th week of pregnancy and determines the due date of delivery as 40 weeks and 0 days. In Patent Document 1 below, an inflection point of the history of heart rate fluctuations is specified, and the predicted due date of the user's (pregnant woman's) child is determined a predetermined number of days after the inflection point.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, even if the due date of delivery can be determined, the onset of labor cannot be accurately predicted, and planned deliveries such as planned cesarean section and planned painless delivery are carried out earlier than necessary, and it has been reported that complications including neonatal respiratory disorders increase. In addition, in order to predict the day of onset of labor, methods of measuring biological data using wearable devices and methods of detecting when labor will come and when it will not come have been published, but these methods have a wide prediction range.

[0005] The present invention has been made in view of the above, and an object thereof is to provide a generation device, a generation method, and a generation program for generating a labor onset prediction model that more accurately predicts the day of onset of labor based on biological data including the heart rate or pulse rate of a pregnant woman.

Means for Solving the Problems

[0006] The present invention relates to a generation device comprising: an acquisition unit that acquires biometric data including heart rate or pulse rate from a user; and a generation unit that uses as training data the biometric data features corresponding to each day from the date of onset of labor to the date of onset of labor, including the day after the date in which the biometric data features were at their lowest, and which are labeled as indicating whether or not the date after a predetermined number of days is the date of onset of labor. The device generates a prediction model that predicts the onset of labor after a predetermined number of days. [Effects of the Invention]

[0007] According to one embodiment, it is possible to generate a model that can predict the onset of labor with greater accuracy. [Brief explanation of the drawing]

[0008] [Figure 1] This figure shows an example of the generation process according to the embodiment. [Figure 2] This figure shows an example (1) of the generation phase of the prediction model according to the embodiment. [Figure 3] This figure shows an example (2) of the generation phase of the prediction model according to the embodiment. [Figure 4] This figure shows the relationship between the generation phase of the prediction model according to the embodiment and the prediction phase of the onset of labor. [Figure 5] This figure shows an example of the prediction phase for the onset of labor according to the embodiment. [Figure 6] This figure shows an example of the display on a terminal device used by a user according to the embodiment. [Figure 7] This figure shows an example of a display on a terminal device used by medical professionals according to the embodiment. [Figure 8] This figure shows an example configuration of a prediction device according to the embodiment. [Figure 9] This is a flowchart showing the procedure for generating the predictive model according to the embodiment. [Figure 10] This flowchart shows the procedure for predicting the onset of labor according to the embodiment. [Figure 11]This is a hardware configuration diagram showing an example of a computer that implements the functions of a prediction device. [Modes for carrying out the invention]

[0009] The following describes in detail, with reference to the drawings, embodiments for implementing the generation apparatus, generation method, and generation program according to the present application. Note that these embodiments do not limit the generation apparatus, generation method, and generation program according to the present application. Furthermore, the same parts are denoted by the same reference numerals in the following embodiments, and redundant descriptions are omitted.

[0010] [1-1. An example of the generation process] First, an example of the generation process according to the embodiment will be described using Figure 1. Figure 1 is a diagram showing an example of the generation process according to the embodiment. The generation process according to the embodiment is performed by the generation system 1 shown in Figure 1. The generation system 1 includes a prediction device 100, which is an example of a generation device, a measurement device 10, and a terminal device 200a. Each device included in the generation system 1 can send and receive data from each other via wireless communication or the like.

[0011] The measurement device 10 is a device that acquires biometric data, including an electrocardiogram or pulse rate. The measurement device 10 is any (wearable) device or apparatus, and may be, for example, a device worn on the user's arm, or a device connected to a predetermined medical device, and acquires biometric data, including the user's electrocardiogram or pulse rate, continuously or periodically. In the following, the user who provides the training data (predetermined data acquired by the measurement device 10) used to generate the prediction model according to the embodiment will be distinguished as "User 15," and the user who is the target of prediction processing by the prediction model will be distinguished as "User 20 (see Figure 5)." When it is not necessary to distinguish between the two, it will simply be referred to as "User." In this embodiment, the user is assumed to be pregnant.

[0012] The measurement device 10 acquires biometric data from the user 15, including an electrocardiogram (ECG) or pulse rate. An electrocardiogram is a graphical record of the electrical activity of the heart. Pulse rate is the number of times the blood vessels in each part of the body beat per minute. If the user's health is normal, the ECG and pulse rate will correspond correctly. The training data used to generate the predictive model according to this embodiment is, for example, the resting heart rate. Resting heart rate indicates how many times the heart beats per minute at rest. The resting heart rate may be calculated based on data measured by the acceleration sensor and electrocardiogram of the measurement device 10, or if the measurement device 10 itself can uniquely acquire the resting heart rate, the acquired data will be used as the resting heart rate. Hereinafter, it will be assumed that the measurement device 10 can uniquely acquire the resting heart rate. Also, the resting heart rate will be simply referred to as "heart rate". The measurement device 10 may also be a wearable device capable of acquiring maximum oxygen uptake (VO2MAX). Since VO2MAX is proportional to the ratio of maximum heart rate to resting heart rate, if a wearable device can acquire VO2MAX, it can also acquire resting heart rate.

[0013] Terminal device 200a is a terminal device used by user 15, such as a smartphone or tablet. For example, an application for controlling the measurement device 10 is installed on terminal device 200a. By using the application running on terminal device 200a, user 15 can control the measurement device 10 and refer to biometric data, including heart rate, measured by the measurement device 10. Terminal device 200a also acquires biometric data, including heart rate, measured by the measurement device 10 and transmits the acquired biometric data to the prediction device 100. Alternatively, the measurement device 10 may directly transmit the biometric data, including heart rate, to the prediction device 100.

[0014] The prediction device 100 is an information processing device that acquires biometric data, including heart rate, measured by the measurement device 10, and, based on the acquired biometric data, captures changes in advance before the onset of labor, generates a labor onset prediction model, and performs processing to predict the onset of labor. For example, the prediction device 100 is a cloud server or the like.

[0015] Although not shown in FIG. 1, it is assumed that there are a large number of users who use the measurement device 10 in addition to the user 15. The prediction device 100 shall acquire biometric data including the heart rate from these numerous users at any time, and use the acquired biometric data as learning data 40 to generate a prediction model described later. That is, the user 15 is a general term for a large number of users who provide biometric data including the heart rate as learning data 40.

[0016] <000009{2]] As described above, the prediction device 100 generates a labor onset prediction model and predicts the onset of labor. Generally, it has been reported that the heart rate of pregnant women decreases in the late stage of pregnancy. However, since the movement of the heart rate after it has decreased varies among pregnant women, it is desirable to focus on the shape of the heart rate waveform. The prediction device 100 provides a visual grasp of the movement of the heart rate toward the day of labor onset, and predicts the onset of labor on an arbitrary day before the day of labor onset using the labor onset prediction model based on the biometric data including the heart rate.

[0017] Therefore, after a certain number of weeks of pregnancy (for example, 28 weeks, and for users who hope for planned delivery, 35 weeks, etc.), the prediction device 100 uses the measurement device 10 to acquire biometric data including the heart rate of the user. The prediction device 100 generates a prediction model based on the feature amounts of the biometric data corresponding to each day from at least the day after the day when the feature amount of the biometric data became the lowest up to the day of labor onset, including the day of labor onset. Regarding the day when the feature amount of the biometric data became the lowest when going back from the day of labor onset, it may be the day when the biometric data (for example, the value of the heart rate) became the lowest when going back from the day of labor onset. And when the predicted value of the onset of labor predicted using the prediction model becomes equal to or higher than a predetermined threshold, for example, in the case of planned delivery, planned delivery is determined including other factors (weeks of pregnancy, degree of dilation of the cervix, etc.).

[0018] The following outline of the prediction model generation process, part of the processing performed by generation system 1, will be explained using Figures 1 to 3. The outline of the labor onset prediction process, also part of the processing performed by generation system 1, will be described later using Figure 5.

[0019] In the example shown in Figure 1, user 15 continuously or periodically acquires biometric data, including heart rate, by wearing the measurement device 10 after a certain gestational week. That is, the measurement device 10 continuously or periodically measures biometric data, including the heart rate of user 15 (step S1). The measurement device 10 transmits the measured biometric data, including heart rate, to the terminal device 200a sequentially or periodically (step S2).

[0020] The terminal device 200a transmits biometric data, including the heart rate measured by the measurement device 10, to the prediction device 100 sequentially or periodically (step S3). Alternatively, the measurement device 10 may directly transmit the biometric data, including the measured heart rate, to the prediction device 100.

[0021] The biometric data, including heart rate, transmitted from the terminal device 200a to the prediction device 100 can be represented as a waveform 30a by plotting the features of the biometric data (e.g., resting heart rate) and the measurement date, as shown in graph 30 in Figure 1. In graph 30, the horizontal axis represents the measurement date, and the vertical axis represents the features of the biometric data, including heart rate. The features of the biometric data are, for example, the moving average of the biometric data. In graph 30, the features of the biometric data are normalized across the most recent data, so the numerical value may exceed 1.

[0022] Graph 30 shows an example of a 28-day moving average of resting heart rate. At the point 30t when the feature quantity falls below a certain level, the predicted value in the predictive model for predicting the onset of labor changes, as will be described later, and the movement up to the onset of labor can be captured.

[0023] The prediction device 100 stores the acquired biometric data, including heart rate, in the training data 40, and generates a prediction model that predicts the onset of labor based on the acquired biometric data, including heart rate (step S4). In other words, the prediction device 100 generates a prediction model by performing machine learning using AI (Artificial Intelligence) based on training data consisting of biometric data, including heart rate, and correct labels related to the biometric data, including heart rate.

[0024] Specifically, the prediction device 100 collects biometric data, including heart rate, from, for example, 28 weeks of gestation onward. In principle, the prediction device 100 collects biometric data corresponding to each day from the onset of labor onward, including at least the day after the day in which the characteristic value of the biometric data was lowest. For users who wish to have a planned delivery, for example, biometric data, including heart rate, from 35 weeks of gestation onward is collected. The biometric data corresponding to each day refers to data obtained, for example, by taking the median of continuously measured data for each day. The biometric data corresponding to each day is not necessarily limited to one data point, but may also be several data points collected by taking the average value every few hours, or multiple data points collected by calculating the average value of data measured every few minutes.

[0025] The prediction device 100 then labels the collected data to indicate whether or not labor began after a predetermined number of days (for example, 3 days). For example, if labor began after 3 days, the prediction device 100 treats this data as positive training data. On the other hand, if labor did not begin after 3 days, the prediction device 100 treats this data as negative training data. To allow for flexibility in prediction, the prediction device 100 may treat not only the day on which labor began, but also the days before and after that day, etc., as positive examples (ground truth data).

[0026] The prediction device 100 can generate sufficient training data for learning by acquiring biometric data, including heart rate, from a large number of users 15. The training data will be explained in detail using Figure 2. Figure 2 is a diagram showing an example of the generation phase of the prediction model according to the embodiment.

[0027] In Graph 50 shown in Figure 2, the horizontal axis represents the measurement date, and the vertical axis represents the features of the biometric data, including heart rate. Graph 50 shows the features 50a of the biometric data, including heart rate, obtained from user 15. The prediction device 100 assigns labels 50b to the features of the biometric data, indicating whether or not labor pains occurred after a predetermined number of days, and generates a set of training data.

[0028] The prediction device 100 then learns based on the training data acquired from the user 15 and generates a prediction model. Specifically, when biometric data including the user's heart rate for multiple days, including the day the user's labor pains began, is input, the prediction device 100 generates a prediction model that predicts the onset of labor after a predetermined number of days. The generation of the prediction model will be explained in detail using Figure 3. Figure 3 is a diagram showing an example of the prediction model generation phase according to the embodiment, following Figure 2. In the graph 60 shown in Figure 3, the horizontal axis is the measurement date, and the vertical axis is the feature quantity of the biometric data including heart rate or the predicted value of the onset of labor. As shown in the graph 60, the prediction device 100 adds a label 60b to the feature quantity 60a of the biometric data including heart rate, indicating whether or not labor has occurred after a predetermined number of days. Then, if the prediction device 100 detects a significant change in the predicted value 60e at the timing 60t when the characteristic quantity 60a of the biometric data, including heart rate, falls below a certain level, it provides a visual representation of the heart rate movement within the range 60c until the onset of labor, and predicts the onset of labor 60d before the onset of labor.

[0029] The prediction model uses biometric data, including heart rate, from, for example, 28 weeks of gestation onwards, but in principle, it is sufficient if it includes the date of onset of labor. For users who wish to have a planned delivery, biometric data including heart rate from, for example, 35 weeks of gestation onwards is used. Furthermore, the prediction device 100 may learn not by using the biometric data including heart rate itself, but by using numerical values ​​that have undergone various preprocessing steps such as moving averages and predetermined normalization. Also, the learning method used by the prediction device 100 is not limited to a specific method, and various known methods (such as regression and clustering) may be used. For example, the prediction device 100 may use any learning method as long as it is possible to extract features based on a certain upward or downward trend in data values ​​as a trend shown by the biometric data including heart rate.

[0030] Here, the relationship between the prediction model generation phase and the labor onset prediction phase, which will be described later, will be explained using Figure 4. Figure 4 is a diagram showing the relationship between the prediction model generation phase and the labor onset prediction phase according to the embodiment. In Figure 4, in the prediction model generation phase, the prediction device 100 preprocesses the biometric data, including heart rate, obtained from the user 15 by interpolating missing values ​​in the measurement data as needed, normalizing the biometric data including heart rate, and calculating the biometric data features. The prediction device 100 then adds labels to the biometric data features indicating whether or not labor will occur after a predetermined number of days to generate a set of training data. The prediction device 100 generates a prediction model using the set of training data. In the labor onset prediction phase, the prediction device 100 performs similar preprocessing on the biometric data, including heart rate, obtained from the user 15, inputs the biometric data features into the trained prediction model, and outputs a score (e.g., probability) indicating whether or not labor will occur.

[0031] Figure 5 illustrates the overview of the prediction process by the generation system 1. Figure 5 shows an example of the prediction phase of the onset of labor according to the embodiment.

[0032] In the labor onset prediction process according to the embodiment, the measurement device 10 continuously or periodically measures biometric data, including the user 20's heart rate (step S11). The measurement device 10 transmits the measured data sequentially or periodically to the terminal device 200a (step S12).

[0033] The terminal device 200a transmits the data measured by the measurement device 10 to the prediction device 100 sequentially or periodically (step S13). Alternatively, the measurement device 10 may directly transmit biometric data, including the measured heart rate, to the prediction device 100.

[0034] The biometric data, including heart rate, transmitted from the terminal device 200a to the prediction device 100 can be represented as a waveform by plotting the heart rate-related features against the measurement date, for example, as shown in graph 70 in Figure 5. In graph 70 in Figure 5, the horizontal axis represents the measurement date, and the vertical axis represents the features of the biometric data, including heart rate, or the predicted value of the onset of labor.

[0035] The prediction device 100 stores biometric data, including heart rate, acquired from the user 20 in the learning data 80. Then, once a predetermined number of days' worth of data has been accumulated, the prediction device 100 predicts whether or not labor will occur after that predetermined number of days (step S14).

[0036] Specifically, the prediction device 100 inputs accumulated biometric data, including heart rate data for multiple days including the day of labor onset, into a prediction model 90 generated by the processing shown in Figures 1 to 3. After the biometric data, including heart rate data, is acquired, the prediction device 100 monitors the feature vector 70a of the biometric data, including heart rate data. If the prediction value 70e changes significantly at timing 70t when the feature vector 70a of the biometric data, including heart rate data, falls below a certain level, the prediction device 100 provides the feature vector 70a of the biometric data, including heart rate data, to the terminal device 200b (described later) for the range 70c until the day of labor onset, so that it can visually grasp the feature vector 70a of the biometric data, including heart rate data. The prediction device 100 predicts the onset of labor 70d before the day of labor onset and outputs a score (e.g., probability) indicating whether or not labor will occur as the prediction result. Alternatively, the prediction device 100 may output the day of labor onset itself instead of a score indicating probability.

[0037] Next, the prediction device 100 transmits the output result to the terminal device 200a (step S15). At this time, the prediction device 100 may control the content displayed in the application on the terminal device 200. For example, the prediction device 100 may control the application to display a message such as, "The probability of labor starting in dd days is XX percent." Specifically, if the prediction model 90 is a 3-day prediction model, the prediction device 100 outputs the probability of labor starting in 3 days.

[0038] The information provided from the prediction device 100 to the terminal device 200a will be explained using Figure 6. Figure 6 is a diagram showing an example of the display on the terminal device used by a user according to the embodiment. As shown in Figure 6, the terminal device 200a displays, for example, "Alert" (200a1), "Data" (200a2), and "Message" (200a3), which show whether there is an abnormality in the measurement data, biometric data including heart rate, a predicted value for the onset of labor, the probability of the onset of labor, etc.

[0039] Furthermore, the prediction device 100 also transmits the output results to the terminal device 200b (step S15). The terminal device 200b is a terminal device used by medical personnel, such as a desktop or tablet terminal. For example, an application for controlling the measurement device 10 is installed on the terminal device 200b. By using the application running on the terminal device 200b, medical personnel can control the measurement device 10 and refer to biometric data, including heart rate, measured by the measurement device 10. The terminal device 200b also receives the predicted results from the prediction device 100.

[0040] The information provided from the prediction device 100 to the terminal device 200b will be explained using Figure 7. Figure 7 shows an example of the display on a terminal device used by a medical professional according to the embodiment. The prediction device 100 outputs prediction results regarding the onset of labor for multiple users who are the target of prediction, using multiple prediction models that predict the onset of labor after multiple predetermined number of days.

[0041] In the example shown in Figure 7, the prediction device 100 monitors multiple pregnant women using multiple models, including a 3-day prediction model that predicts the onset of labor in 3 days and a 7-day prediction model that predicts the onset of labor in 7 days. Therefore, as shown in Figure 7, the terminal device 200b displays, for example, multiple users, each corresponding to a different ID, with different displays 200b2 indicating whether or not labor will begin within the corresponding predetermined days. In other words, the terminal device 200b displays a real-time prediction 200b1 of the onset of labor for multiple users.

[0042] For example, for a user with "ID" "001", a real-time prediction is made that labor will begin within 3 days, and the expected delivery date is displayed as August 20, 2024. Similarly, for a user with "ID" "002", a real-time prediction is made that labor will begin within 7 days, and the expected delivery date is displayed as August 22, 2024. And for a user with "ID" "005", a real-time prediction is made that labor will not begin within 7 days, and the expected delivery date is displayed as August 30, 2024.

[0043] In this way, the prediction device 100 acquires biometric data, including heart rate, from the user, generates a model to predict the onset of labor based on the acquired biometric data, and uses this model to predict the onset date of labor with greater accuracy.

[0044] [1-2. Configuration of the prediction device according to the embodiment] Next, the configuration of the prediction device 100 that performs the labor onset prediction processing according to the embodiment will be described. Figure 8 is a diagram showing an example of the configuration of the prediction device 100 according to the embodiment.

[0045] As shown in Figure 8, the prediction device 100 includes a communication unit 110, a storage unit 120, and a control unit 130. The prediction device 100 may also include an input unit (e.g., a keyboard or mouse) for receiving various operations from an administrator or other person managing the prediction device 100, and a display unit (e.g., a liquid crystal display) for displaying various information.

[0046] The communication unit 110 is implemented, for example, by a network interface controller. The communication unit 110 is connected to a network N (e.g., the Internet) by wire or wireless connection and transmits and receives information to and from the measurement device 10, terminal device 200, etc. via the network N. For example, the communication unit 110 may transmit and receive information using communication standards or technologies such as Wi-Fi®, SIM (Subscriber Identity Module), or LPWA (Low Power Wide Area).

[0047] The storage unit 120 is implemented by, for example, semiconductor memory elements such as RAM (Random Access Memory) and flash memory, or by storage devices such as hard disks and optical discs. The storage unit 120 has a measurement data storage unit 121 and a model storage unit 122.

[0048] The measurement data storage unit 121 stores measurement data acquired from the user wearing the measurement device 10. Specifically, the measurement data storage unit 121 stores biometric data including heart rate, which is training data used to train the prediction model, and biometric data including heart rate, which is input to the prediction model during the prediction process.

[0049] The model storage unit 122 stores the prediction models that the prediction device 100 uses for prediction processing. The model storage unit 122 may also store multiple prediction models that predict the onset of labor after a predetermined number of days.

[0050] The control unit 130 is implemented, for example, by a CPU (Central Processing Unit), MPU (Micro Processing Unit), GPU (Graphics Processing Unit), etc., which executes a program stored inside the prediction device 100 using RAM or the like as a working area. The control unit 130 is also a controller and is implemented, for example, by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or FPGA (Field Programmable Gate Array).

[0051] As shown in Figure 8, the control unit 130 includes an acquisition unit 131, a generation unit 132, a preprocessing unit 133, a prediction unit 134, and a supply unit 135.

[0052] The acquisition unit 131 acquires various types of information. For example, the acquisition unit 131 acquires biometric data, including heart rate or pulse rate, from a user using the measurement device 10. Specifically, the acquisition unit 131 controls a program (app) installed on the terminal device 200a and acquires biometric data, including heart rate or pulse rate, from the terminal device 200a at the time the biometric data, including heart rate or pulse rate, is acquired by the terminal device 200a, or at a fixed time each day. Alternatively, the acquisition unit 131 may acquire biometric data, including heart rate or pulse rate, from the measurement device 10 at the time the biometric data, including heart rate or pulse rate, is acquired by the measurement device 10, or at a fixed time each day.

[0053] Furthermore, the acquisition unit 131 may acquire not only biometric data including heart rate or pulse rate, but also any other physical information that is expected to affect the onset of labor, such as the user's age, height, and weight. The acquisition unit 131 acquires this information by receiving input from the user, for example, through user registration in the app. In addition, if the measurement device 10 is capable of measuring biometric information other than biometric data including heart rate or pulse rate, the acquisition unit 131 may acquire such biometric information together with the biometric data including heart rate or pulse rate.

[0054] The generation unit 132 uses as training data the features of the biometric data including heart rate or pulse rate corresponding to each day from the date of onset of labor to the date of onset of labor, including the day on which the features of the biometric data including heart rate or pulse rate were at their lowest, and labels whether the date after a predetermined number of days is the date of onset of labor. This data is used to generate a predictive model that predicts the onset of labor after a predetermined number of days. Note that "backward from the date of onset of labor" refers to the period from the day on which observation (acquisition) of the user's heart rate data began to the date of onset of labor. The day on which the features of the biometric data including heart rate or pulse rate were at their lowest is the day on which the features of the biometric data including heart rate or pulse rate reached their lowest value during the period from the day on which acquisition of biometric data including heart rate or pulse rate began to the date of onset of labor. The day on which the features of the biometric data were at their lowest, backward from the date of onset of labor, may also be the day on which the biometric data including heart rate or pulse rate (for example, a specific observed value of heart rate) was at its lowest, backward from the date of onset of labor.

[0055] The generation unit 132 uses the acquired biometric data, including heart rate or pulse rate, and ground truth data indicating whether labor began a predetermined number of days after the biometric data was measured (for example, 3 days later), as training data to generate a predictive model that predicts the onset of labor from the biometric data, including heart rate.

[0056] The generation unit 132 may generate a predictive model by assigning different weights to training data that shows labor pains occurred after a predetermined number of days (positive examples) and training data that shows labor pains did not occur after a predetermined number of days (negative examples). The generation unit 132 may use various known methods as the weighting method.

[0057] The preprocessing unit 133 performs predetermined preprocessing on the biometric data, including heart rate, acquired by the acquisition unit 131, in order to use it as training data.

[0058] For example, if there is missing measurement data in the biometric data, including heart rate, the preprocessing unit 133 interpolates the missing measurement data using measurement data measured on the days before and after the missing measurement data.

[0059] Furthermore, the preprocessing unit 133 may perform normalization processing as appropriate to make the data easier to handle in machine learning.

[0060] The prediction unit 134 uses a prediction model to predict the onset of labor for user 20 when it obtains biometric data, including heart rate, for a predetermined number of days (for example, 8 weeks from 28 weeks of pregnancy). Specifically, the prediction unit 134 inputs the biometric data, including heart rate, obtained from user 20 into the prediction model and outputs a score indicating whether or not labor will begin after the predetermined number of days.

[0061] The provisioning unit 135 provides the user 20, who is the target of the prediction, with prediction results regarding the onset of labor, which are predicted using multiple prediction models that predict the onset of labor after a predetermined number of days by the prediction unit 134. For example, the provisioning unit 135 transmits the results predicted by the prediction unit 134 to the terminal device 200a. At this time, if the provisioning unit 135 observes signs such as the onset of labor being imminent in the prediction results, it may display (notify) the user 20 with more emphasis than usual.

[0062] Furthermore, the provisioning unit 135 may send advice to the user based on the predicted onset of labor by the prediction unit 134. For example, the provisioning unit 135 may send advice regarding the estimated physical condition of the user 20 and the timing of hospitalization.

[0063] The service provider 135 may send different advice to each user 20. For example, the service provider 135 may refer to past history (such as registration by the user 20) and send advice regarding the duration of labor according to the number of births.

[0064] Furthermore, the provisioning unit 135 transmits to the terminal device 200b a real-time prediction of the onset of labor for multiple users who are the target of the prediction, using multiple prediction models that predict the onset of labor after multiple predetermined number of days by the prediction unit 134. The real-time prediction is a prediction of whether or not labor will start within a predetermined number of days (for example, 3 days later or 7 days later). The provisioning unit 135 also transmits prediction information such as the expected date of delivery to the terminal device 200b.

[0065] [1-3. Procedure for generating a prediction model and predicting the onset of labor according to the embodiment] The processing flow of the prediction device 100 described above will be explained using Figures 9 and 10. Figure 9 is a flowchart showing the procedure for generating a prediction model according to the embodiment.

[0066] As shown in Figure 9, the prediction device 100 acquires biometric data from the user, including heart rate or pulse rate, which is used to generate the prediction model (step S101).

[0067] Next, the prediction device 100 performs data preprocessing on the biological data, including heart rate or pulse rate, such as interpolation of missing data and normalization (step S102).

[0068] Next, the prediction device 100 labels the pre-processed data to indicate whether or not labor pains occurred after a predetermined number of days. If labor pains occurred after the predetermined number of days, the data is used as positive training data; if labor pains did not occur, the data is used as negative training data, and training data is generated (step S103).

[0069] Then, the prediction device 100 performs the learning process described above and generates a prediction model (step S104). The prediction device 100 stores the generated prediction model in the storage unit 120 and terminates the process.

[0070] Next, the flow of the prediction process will be explained using Figure 10. Figure 10 is a flowchart showing the procedure for predicting the onset of labor according to the embodiment.

[0071] As shown in Figure 10, the prediction device 100 acquires biometric data, including heart rate or pulse rate, from the user who is the subject of the prediction (step S201). The prediction device 100 then determines whether the data including heart rate or pulse rate necessary for prediction has been collected (step S202). If the biometric data necessary for prediction has not been collected (step S202; No), the prediction device 100 waits until it can acquire more biometric data including heart rate or pulse rate from the user who is the subject of the prediction.

[0072] On the other hand, once the necessary biometric data for prediction has been collected (Step S202; Yes), the prediction device 100 preprocesses the acquired biometric data, including heart rate or pulse rate, and inputs it into the prediction model (Step S203). Then, the prediction device 100 predicts the onset of labor based on the output score (Step S204).

[0073] Subsequently, the prediction device 100 provides prediction results and advice to the user and medical professionals who are the subjects of the prediction (step S205).

[0074] [2. Modified Examples of Embodiments] In the above embodiment, the terminal device 200a according to the present application was shown as an example of acquiring biometric data including heart rate measured by the measurement device 10. However, the terminal device 200a does not necessarily have to acquire biometric data including heart rate from the measurement device 10. For example, the terminal device 200a may directly accept input from the user regarding biometric data including resting heart rate acquired by the user using a general smart device equipped with a heart rate sensor, and treat the accepted data as input data or learning data.

[0075] Furthermore, in the embodiment, an example was shown in which the terminal device 200a or the measurement device 10 transmits measurement data to the prediction device 100, and the prediction device 100 makes a prediction of the onset of labor. However, if the terminal device 200a has a prediction model, the terminal device 200a may predict the onset of labor for the user based on biometric data, including heart rate, measured by the measurement device 10. In other words, the processing performed by the prediction device 100 in the embodiment may be performed by the terminal device 200a.

[0076] Furthermore, in this embodiment, the prediction device 100 is shown to have a storage unit 120 that holds measurement data and prediction models, but this information may also be held by an external storage device other than the prediction device 100.

[0077] [3. Other Embodiments] The processing according to the above-described embodiment may be carried out in various other forms besides those described above.

[0078] For example, among the processes described in the above embodiments, all or part of the processes described as being performed automatically can be performed manually, or all or part of the processes described as being performed manually can be performed automatically by known methods. In addition, the processing procedures, specific names, and information including various data and parameters shown in the above document and drawings can be changed at will unless otherwise specified. For example, the various information shown in each figure is not limited to the information shown.

[0079] Furthermore, the components of each illustrated device are functionally conceptual and do not necessarily need to be physically configured as shown. In other words, the specific forms of distribution and integration of each device are not limited to those shown, and all or part of them can be functionally or physically distributed and integrated in any unit according to various loads and usage conditions.

[0080] Furthermore, the embodiments and modifications described above can be combined as appropriate, provided that the processing content is not inconsistent.

[0081] Furthermore, the effects described herein are merely illustrative and not limiting; other effects may also occur.

[0082] [4. Effects of the generating device relating to this application] As described above, the generation device (terminal device 200a and prediction device 100) according to the present invention comprises an acquisition unit 131 that acquires biometric data including heart rate or pulse rate from a user, and a generation unit 132 that uses as training data the biometric data features corresponding to each day from the date of onset of labor to the date of onset of labor, including the day after the date in which the biometric data features were at their lowest, and which are labeled as indicating whether or not the date after a predetermined number of days is the date of onset of labor. The generation device generates a prediction model that predicts the onset of labor after the predetermined number of days.

[0083] The generation apparatus according to the present invention further comprises a prediction unit 134 that uses a prediction model to predict the onset of labor for a user when biometric data is acquired from the user to be predicted by the acquisition unit 131, and a provision unit 135 that provides information related to the onset of labor predicted by the prediction unit 134.

[0084] The generation apparatus according to the present invention further comprises a preprocessing unit 133 that performs predetermined preprocessing on the biological data acquired by the acquisition unit 131 for use as training data, the preprocessing unit 133 normalizes the moving average of the features of the biological data, and the generation unit 132 generates the prediction model using the data normalized by the preprocessing unit 133 as the training data.

[0085] If there is missing measurement data in the biological data, the preprocessing unit 133 interpolates the missing measurement data using measurement data measured on dates before and after the missing measurement data.

[0086] The provision unit 135 provides the terminal device owned by a medical professional with information regarding the onset of labor, which has been predicted by the prediction unit 134 using the prediction model for multiple users who are the target of the prediction.

[0087] The provision unit 135 provides information regarding the onset of labor, which has been predicted by the prediction unit 134 using multiple prediction models that predict the onset of labor after multiple predetermined number of days for multiple users who are the target of prediction, to terminal devices owned by medical personnel.

[0088] By performing any or a combination of the above-described processes, the generation device according to the present invention can generate a model that predicts the onset of labor with greater accuracy.

[0089] [5. Hardware Configuration] The information devices such as the prediction device 100 and terminal device 200 according to the embodiments described above are realized by a computer 1000 having a configuration such as that shown in Figure 11. The following explanation will use the prediction device 100 according to the embodiments as an example. Figure 11 is a hardware configuration diagram showing an example of a computer 1000 that realizes the functions of the prediction device 100. The computer 1000 has a CPU 1100, RAM 1200, ROM (Read Only Memory) 1300, HDD (Hard Disk Drive) 1400, communication interface 1500, and input / output interface 1600. The various parts of the computer 1000 are connected by a bus 1050.

[0090] The CPU 1100 operates based on programs stored in the ROM 1300 or HDD 1400, and controls various parts. For example, the CPU 1100 loads the programs stored in the ROM 1300 or HDD 1400 into the RAM 1200 and executes processing corresponding to the various programs.

[0091] ROM1300 stores boot programs such as the BIOS (Basic Input Output System) executed by CPU1100 when computer 1000 starts up, as well as programs that depend on the computer 1000's hardware.

[0092] The HDD1400 is a computer-readable recording medium that non-temporarily stores programs executed by the CPU1100 and data used by such programs. Specifically, the HDD1400 is a recording medium that stores a program that performs the labor onset prediction processing according to the present invention, which is an example of program data 1450.

[0093] The communication interface 1500 is an interface for the computer 1000 to connect to an external network 1550 (e.g., the Internet). For example, the CPU 1100 can receive data from other devices or transmit data it generates to other devices via the communication interface 1500.

[0094] The input / output interface 1600 is an interface for connecting the input / output device 1650 and the computer 1000. For example, the CPU 1100 receives data from input devices such as a keyboard or mouse via the input / output interface 1600. The CPU 1100 also transmits data to output devices such as a display, speaker, or printer via the input / output interface 1600. The input / output interface 1600 may also function as a media interface for reading programs recorded on a predetermined recording medium (media). Examples of media include optical recording media such as DVDs (Digital Versatile Discs) and PDs (Phase Change Rewritable Disks), magneto-optical recording media such as MOs (Magneto-Optical Disks), tape media, magnetic recording media, or semiconductor memory.

[0095] For example, when the computer 1000 functions as a prediction device 100 according to the embodiment, the CPU 1100 of the computer 1000 realizes functions such as the control unit 130 by executing a labor onset prediction processing program loaded on the RAM 1200. The HDD 1400 stores the program that executes the labor onset prediction processing according to the present invention, as well as data in the storage unit 120. The CPU 1100 reads and executes the program data 1450 from the HDD 1400, but as another example, these programs may be obtained from other devices via an external network 1550.

[0096] Although embodiments of the present application have been described in detail based on the drawings, these are illustrative examples, and the present invention can be implemented in various other forms with modifications and improvements based on the knowledge of those skilled in the art, starting with the embodiments described in the disclosure section of the invention. [Explanation of symbols]

[0097] 1. Generation System 10 Measuring Devices 100 Prediction Devices 110 Communications Department 120 Storage section 121 Measurement data storage unit 122 Model Memory Unit 130 Control Unit 131 Acquisition Department 132 Generation part 133 Pre-processing section 134 Prediction Section 135 Provision Department

Claims

1. An acquisition unit that acquires biometric data including heart rate or pulse rate from the user, A generation unit generates a predictive model that predicts the onset of labor after a predetermined number of days, using as training data the features of the biological data corresponding to each day up to the onset of labor, including the day after the day in which the features of the biological data were at their lowest, and labeling whether or not the day after a predetermined number of days is the day of labor. A generating apparatus characterized by comprising the following features.

2. When the acquisition unit acquires the biometric data from the user to be predicted, the prediction unit uses the prediction model to predict when labor pains will begin for the user to be predicted, A providing unit that provides information related to the onset of labor pains predicted by the prediction unit, The generating apparatus according to claim 1, further comprising the features described above.

3. The system further includes a preprocessing unit that performs predetermined preprocessing on the biological data acquired by the acquisition unit in order to use it as learning data. The aforementioned pre-processing unit, The moving average of the features of the aforementioned biometric data is normalized, The generating unit is The data normalized by the preprocessing unit is used as the training data to generate the prediction model. The generating apparatus according to feature 2.

4. The aforementioned pre-processing unit, If there are missing measurement data points among the aforementioned biological data, the missing data points will be interpolated using measurement data taken on dates before and after the missing measurement data points. The generating apparatus according to feature 3.

5. The aforementioned supply unit is, The prediction unit provides information regarding the onset of labor, predicted using the prediction model, for multiple users who are the target of the prediction, to terminal devices owned by medical personnel. The generating apparatus according to feature 2.

6. The aforementioned supply unit is, The prediction unit provides information regarding the onset of labor, predicted using multiple prediction models that predict the onset of labor after multiple predetermined days for multiple users who are the target of prediction, to a terminal device owned by a medical professional. The generating apparatus according to feature 2.

7. A generation method performed by a computer, The process involves acquiring biometric data, including heart rate or pulse rate, from the user. A generation step of generating a predictive model that predicts the onset of labor after a predetermined number of days, using as training data the features of the biological data corresponding to each day from the date of onset of labor to the date of onset of labor, including the day after the date in which the features of the biological data were at their lowest, and labeling whether or not the date after a predetermined number of days is the date of onset of labor. A method for generating a product, characterized by including the following:

8. Procedure for obtaining biometric data, including heart rate or pulse rate, from a user, A generation procedure for generating a predictive model that predicts the onset of labor after a predetermined number of days, using as training data the features of the biometric data corresponding to each day from the date of onset of labor to the date of onset of labor, including the day after the date in which the features of the biometric data were at their lowest, and which are labeled as whether or not the date after a predetermined number of days is the date of labor. A generation program characterized by causing a computer to execute it.

Images