Information processing device, information processing method, and program

Abstract

To provide an information processing device, an information processing method, and a program capable of reducing learning time of a learning model and at the same time predicting an appropriate instruction by predicting sleeping time of a subject to be observed.SOLUTION: An information processing device comprises a first learning model machine-learned so as to predict sleeping time of a subject on the basis of recorded data regarding the subject who becomes an observation object of a worker created by the worker who is engaged in medical care or nursing, and a second learning model machine-learned so as to predict an instruction to the subject on the basis of sleeping time of the subject predicted by using the first learning model.SELECTED DRAWING: Figure 4

Description

[Technical Field]

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

[0002] In medical and nursing care settings, multiple medical and nursing staff are involved in the care and daily living of patients and care recipients. Each staff member observes and diagnoses the patient's or care recipient's condition. A single staff member does not observe the patient or care recipient continuously; rather, multiple staff members from various professions visit and observe the patient or care recipient at different times and intervals. Therefore, in order to share information about the patient or care recipient, each staff member registers and views the results of their patient observations in an electronic medical record system, a nursing and care record system, or a social networking service (SNS).

[0003] As a technology that utilizes such observation results of a patient or care recipient, a technology has been disclosed that uses machine learning to predict diseases that the patient or care recipient may develop in the future or events that the patient or care recipient may experience (e.g., Patent Document 1). Another technology has been disclosed that uses a load sensor or the like attached to a bed to detect at least a portion of the subject's biological status, such as the subject's heart rate, pulse rate, respiratory rate, body temperature, blood pressure, blood oxygen saturation, weight, sleep, and wakefulness (e.g., Patent Document 2). Another technology has been disclosed that acquires information about the condition of the care recipient based on input from a Doppler sensor or camera and references the subject's sleep status over multiple past days from diary data (e.g., Patent Document 3). Another technology has been disclosed that can grasp sleep status using a camera (e.g., Patent Document 4). Another technology has been disclosed that uses a camera-equipped PDA to capture an overall image of the subject before eating and images of leftover food, etc., and transmits them to a server, enabling the acquired calorie data to be effectively utilized (see Patent Document 5). Summary of the Invention [Problem to be solved by the invention]

[0004] However, the technology described in Patent Document 1 can predict a subject's real-time risk and disease onset using a learning model based on health information, activity information, vital signs sensing information, plan information, record information, event information, etc., but has the problem that it only uses one learning model for the machine learning model, and therefore requires a long learning time for sufficient learning. Furthermore, the technology described in Patent Document 2 can measure sleep using a load sensor attached to the bed, but has the problem of not being able to predict any instructions based on the care recipient's sleep time. Furthermore, the technology described in Patent Document 3 can refer to the subject's sleep status for multiple past days from diary data, but has the problem of not being able to predict any instructions based on the care recipient's sleep time. Furthermore, the technology described in Patent Document 4 can grasp the sleep state using a camera, but has the problem of not being able to predict any instructions based on the care recipient's sleep time. Furthermore, the technology described in Patent Document 5 can effectively use calorie data acquired by a camera-equipped PDA, but has the problem of not being able to predict any instructions based on the care recipient's sleep time.

[0005] The present invention has been made in consideration of the above, and aims to provide an information processing device, an information processing method, and a program that can predict appropriate instructions by predicting the sleep time of a subject being observed while reducing the learning time of a learning model. [Means for solving the problem]

[0006] In order to solve the above-mentioned problems and achieve the objectives, the present invention is characterized by comprising a first learning model trained by machine learning to predict the sleeping time of a subject based on recorded data about the subject that is the subject of observation by a medical or nursing care worker, and a second learning model trained by machine learning to predict instructions to the subject based on the sleeping time of the subject predicted using the first learning model. [Effects of the Invention]

[0007] According to the present invention, it is possible to predict appropriate instructions by predicting the sleep duration of a subject being observed while reducing the learning time of a learning model. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram illustrating an example of the overall configuration of an information processing system according to the first embodiment. [Figure 2] FIG. 2 is a diagram illustrating an example of a hardware configuration of the information processing apparatus according to the first embodiment. [Figure 3] FIG. 3 is a diagram illustrating an example of a hardware configuration of the employee terminal according to the first embodiment. [Figure 4] FIG. 4 is a diagram illustrating an example of the functional block configuration of the information processing device according to the first embodiment. [Figure 5] FIG. 5 is a diagram showing an example of recorded data in the information processing system according to the first embodiment. [Figure 6] FIG. 6 is a flowchart showing an example of the flow of the sleep time learning process of the information processing device according to the first embodiment. [Figure 7] FIG. 7 is a flowchart showing an example of the flow of the warning learning process of the information processing device according to the first embodiment. [Figure 8] FIG. 8 is a flowchart showing an example of the flow of a sleep time prediction process of the information processing device according to the first embodiment. [Figure 9]FIG. 9 is a diagram illustrating an example of a sleep time prediction history of the information processing device according to the first embodiment. [Figure 10] FIG. 10 is a flowchart showing an example of the flow of the warning prediction process of the information processing device according to the first embodiment. [Figure 11] FIG. 11 is a diagram showing an example of the text of an email displayed on the employee terminal according to the first embodiment. [Figure 12] FIG. 12 is a diagram showing an example of recorded data in the information processing system according to the second embodiment. [Figure 13] FIG. 13 is a flowchart showing an example of the flow of a menu learning process of the information processing device according to the second embodiment. [Figure 14] FIG. 14 is a flowchart showing an example of the flow of a menu prediction process performed by the information processing device according to the second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, with reference to the drawings, embodiments of an information processing device, an information processing method, and a program according to the present invention will be described in detail. Furthermore, the present invention is not limited to the following embodiments, and the components in the following embodiments include those that would be easily conceived by a person skilled in the art, those that are substantially the same, and those that are within the scope of what is called equivalents. Furthermore, various omissions, substitutions, modifications, and combinations of the components can be made without departing from the spirit of the following embodiments.

[0010] [First embodiment] (Overall configuration of information processing system) 1 is a diagram showing an example of the overall configuration of an information processing system according to the first embodiment. The overall configuration of the information processing system 1 according to this embodiment will be described with reference to FIG.

[0011] The information processing system 1 shown in Fig. 1 is a system for predicting a sleep duration based on observation results of a subject (hereinafter simply referred to as a subject), such as a care recipient or a patient, and predicting instructions for the subject based on the predicted sleep duration. As shown in Fig. 1, the information processing system 1 includes an information processing device 10 and staff terminals 20a and 20b. The information processing device 10 and staff terminals 20a and 20b are capable of data communication with each other via a network N, such as a LAN (Local Area Network) or the Internet.

[0012] The number of staff terminals 20a, 20b is not limited to two, and may be any other number. Furthermore, when referring to any staff terminal or collectively referring to these staff terminals 20a, 20b, etc., they will be simply referred to as "staff terminal 20."

[0013] The information processing device 10 is a computer such as a PC (Personal Computer), workstation, or server device that predicts the sleep duration of a subject from record data such as a diary or medical chart registered from the staff terminal 20 and predicts instructions for the subject from the predicted sleep duration. Note that the information processing device 10 is not limited to being realized by a single computer, but may be realized by distributed processing using multiple computers. Furthermore, the information processing device 10 may be realized by a computer in a cloud environment or a computer in an on-premise environment.

[0014] The staff terminal 20 is a PC, mobile phone, smartphone, tablet terminal, PDA (Personal Digital Assistant), wearable terminal, or the like, through which staff such as medical professionals and caregivers (hereinafter, sometimes simply referred to as staff) who provide nursing care or care to a subject input record data such as a journal or medical chart containing the subject's observation results. The journals and the like included in the record data are not limited to daily records, but may be hourly records or monthly records, for example. The staff terminal 20 transmits the input record data to the information processing device 10 via the network N, and the record data is registered in the information processing device 10. In this embodiment, the staff terminal 20 is described as being a smartphone. The staff terminal 20 may also be a projector, an electronic whiteboard, digital signage, a head-up display device, industrial machinery, an imaging device, a sound collection device, a medical device, a network home appliance, a connected car, or the like.

[0015] (Hardware configuration of information processing device) 2 is a diagram showing an example of the hardware configuration of the information processing device according to the first embodiment. The hardware configuration of the information processing device 10 according to this embodiment will be described with reference to FIG.

[0016] As shown in FIG. 2, the information processing device 10 includes a CPU (Central Processing Unit) 701, a ROM (Read Only Memory) 702, a RAM (Random Access Memory) 703, an auxiliary storage device 705, a media drive 707, a display 708, a network I / F 709, a keyboard 711, a mouse 712, and a DVD (Digital Versatile Disc) drive 714.

[0017] The CPU 701 is a computing device that controls the overall operation of the information processing device 10. The ROM 702 is a non-volatile storage device that stores programs for the information processing device 10. The RAM 703 is a volatile storage device that is used as a work area for the CPU 701.

[0018] The auxiliary storage device 705 is a storage device such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive) that stores various data, programs, etc. The media drive 707 is a device that controls reading and writing of data from and to a recording medium 706 such as a flash memory under the control of the CPU 701.

[0019] The display 708 is a display device configured with a liquid crystal or organic EL (Electro-Luminescence) display, etc., that displays various information such as a cursor, a menu, a window, characters, or an image.

[0020] The network I / F 709 is an interface for communicating data with external devices such as the employee terminal 20 via the network N. The network I / F 709 is, for example, a NIC (Network Interface Card) that supports Ethernet (registered trademark) and is capable of wired or wireless communication in accordance with TCP (Transmission Control Protocol) / IP (Internet Protocol) or the like.

[0021] The keyboard 711 is an input device for selecting letters, numbers, and various instructions, moving the cursor, etc. The mouse 712 is an input device for selecting and executing various instructions, selecting a processing target, moving the cursor, etc.

[0022] The DVD drive 714 is a device that controls reading and writing of data from and to a DVD 713 such as a DVD-ROM or a DVD-R (Digital Versatile Disk Recordable) as an example of a removable storage medium.

[0023] The above-mentioned CPU 701, ROM 702, RAM 703, auxiliary storage device 705, media drive 707, display 708, network I / F 709, keyboard 711, mouse 712 and DVD drive 714 are communicatively connected to each other via a bus 710 such as an address bus and a data bus.

[0024] Note that the hardware configuration of the information processing device 10 shown in Fig. 2 is an example, and does not necessarily include all of the components shown in Fig. 2, or may include other components. Furthermore, the information processing device 10 is not limited to being configured as a single information processing device as shown in Fig. 2, and may be configured as multiple information processing devices as described above.

[0025] (Hardware configuration of employee terminals) 3 is a diagram showing an example of the hardware configuration of the staff terminal 20 according to the first embodiment. The hardware configuration of the staff terminal 20 according to the present embodiment will be described with reference to FIG.

[0026] As shown in FIG. 3, the employee terminal 20 includes a CPU 801, a ROM 802, a RAM 803, an EEPROM (Electrically Erasable Programmable Read Only Memory) 804, an imaging unit 805, an imaging I / F 806, an acceleration / direction sensor 807, and a GNSS (Global Navigation Satellite System) receiving unit 808.

[0027] The CPU 801 is a computing device that controls the overall operation of the staff terminal 20. The ROM 802 is a non-volatile storage device that stores programs used to drive the CPU 801, such as an IPL (Initial Program Loader). The RAM 803 is a volatile storage device that is used as a work area for the CPU 801. The EEPROM 804 is a non-volatile storage device that stores programs and various data.

[0028] The imaging unit 805 is a built-in imaging device (camera) that captures an image of a subject using an image sensor such as a CMOS (Complementary Metal Oxide Semiconductor) to obtain image data under the control of the CPU 801. Note that instead of a CMOS image sensor, an image sensor such as a CCD (Charge Coupled Device) may also be used. The imaging I / F 806 is an interface for controlling the driving of the imaging unit 805.

[0029] The acceleration / direction sensor 807 is one of various sensors such as an electronic magnetic compass that detects geomagnetism, a gyrocompass, an acceleration sensor, and the like.

[0030] The GNSS receiver 808 is a receiving device that receives positioning signals from positioning satellites, such as GPS (Global Positioning System) signals from GPS satellites.

[0031] As shown in FIG. 3, the employee terminal 20 further includes a long-distance communication circuit 810, an antenna 810a, a short-distance communication circuit 811, an antenna 811a, a microphone 812, a speaker 813, a sound input / output I / F 814, a display 815, an external device connection I / F 816, a vibrator 817, and a touch panel 818.

[0032] The long-distance communication circuit 810 is a communication circuit that performs wireless communication with other devices via a network N through an antenna 810a in accordance with standards such as Wi-Fi (registered trademark).

[0033] The short-distance communication circuit 811 is a communication circuit that performs short-distance wireless communication with other devices via an antenna 811a in accordance with standards such as NFC (Near Field Communication) or Bluetooth (registered trademark).

[0034] The microphone 812 is a built-in sound collecting device that converts sound into an electrical signal. The speaker 813 is a built-in acoustic device that converts the electrical signal into physical vibrations and outputs sound such as music or voice. The sound input / output I / F 814 is an interface that processes input and output of sound signals between the microphone 812 and the speaker 813 under the control of the CPU 801. The microphone 812 and the speaker 813 may be wirelessly connected headsets.

[0035] The display 815 is a display device such as a liquid crystal display or an organic EL display that displays an image of a subject, various icons, etc. The external device connection I / F 816 is an interface conforming to standards such as USB (Universal Serial Bus) for connecting various external devices.

[0036] The vibrator 817 is a device that generates physical vibrations under the control of the CPU 801 .

[0037] The touch panel 818 is an input device that allows the worker to perform various functions of the worker terminal 20 by touching the display 815 .

[0038] The above-mentioned CPU 801, ROM 802, RAM 803, EEPROM 804, imaging I / F 806, acceleration / direction sensor 807, GNSS receiver 808, long-distance communication circuit 810, short-distance communication circuit 811, sound input / output I / F 814, display 815, external device connection I / F 816, vibrator 817, and touch panel 818 are connected to each other so as to be able to communicate with each other via bus lines 809 such as an address bus and a data bus.

[0039] The hardware configuration of the staff terminal 20 shown in FIG. 3 is an example, and it is not necessary to include all of the components, and other components may also be included.

[0040] (Configuration and operation of functional blocks of information processing device) Fig. 4 is a diagram showing an example of the configuration of functional blocks of the information processing device according to the first embodiment. Fig. 5 is a diagram showing an example of recorded data in the information processing system according to the first embodiment. The configuration and operation of the functional blocks of the information processing device 10 according to this embodiment will be described with reference to Figs. 4 and 5.

[0041] As shown in FIG. 4, the information processing device 10 includes a communication unit 101, a morphological analysis unit 102, a sleep time learning unit 103 (first learning unit), an instruction learning unit 104 (second learning unit), a sleep time prediction unit 105 (first prediction unit), an instruction prediction unit 106 (second prediction unit), an email generation unit 107, a notification unit 108, and a memory unit 110.

[0042] The communication unit 101 is a functional unit that performs data communication with the staff terminal 20 and the like via the network I / F 709. For example, the communication unit 101 receives record data input by a staff member from the staff terminal 20 via the network N, and registers the data in a record data DB 111 of the storage unit 110, which will be described later.

[0043] The morphological analysis unit 102 is a functional unit that breaks down text data such as report contents in the record data registered in the record data DB 111 of the storage unit 110 and the record data included in the learning data stored in the learning data DB 115 into minimum units (morphemes) such as words through morphological analysis processing. Note that the function of the morphological analysis unit 102 may be realized by using existing morphological analysis software.

[0044] The sleep time learning unit 103 is a functional unit that generates a learning model for predicting sleep time through machine learning using the learning data stored in the learning data DB 115 of the storage unit 110. Specifically, the sleep time learning unit 103 generates the learning model through machine learning using the results of morphological analysis processing by the morphological analysis unit 102 on the learning data stored in the learning data DB 115. That is, machine learning is performed on the record data shown in FIG. 5 (described later) based on sleep-related words and phrases and date and time decomposed by the morphological analysis processing by the morphological analysis unit 102, thereby generating a learning model with improved accuracy in predicting sleep time. Here, the learning data stored in the learning data DB 115 used by the sleep time learning unit 103 is, for example, data in which the record data registered in the record data DB 111 is associated with the subject's actual sleeping time as a label after the record data is registered by the practitioner. Furthermore, the machine learning may use any of a neural network such as a recurrent neural network (RNN), a support vector machine (SVM), a conditional random fields (CRF), etc. Then, the sleep time learning unit 103 stores the generated learning model in the storage unit 110 as a sleep time learning model 116 (first learning model).

[0045] The instruction learning unit 104 is a functional unit that generates a learning model for predicting warnings (examples of instructions) for a subject, such as a fall, aspiration, or incontinence, by machine learning using the learning data stored in the learning data DB 115 of the storage unit 110. The warnings here refer to warnings for the subject's behavior, such as a fall, aspiration, or incontinence. Specifically, the instruction learning unit 104 generates the learning model by machine learning using the results of morphological analysis performed by the morphological analysis unit 102 on the learning data stored in the learning data DB 115. The learning data stored in the learning data DB 115 used by the instruction learning unit 104 is, for example, record data registered in the record data DB 111, and data in which a report date and time and a predicted sleep time predicted by the sleep time prediction unit 105 (described later) and stored in the sleep time prediction history DB 112 of the storage unit 110 are associated as labels with events, such as a fall, aspiration, or incontinence, that occurred in the subject after the record data was registered by the staff member. Note that the learning data does not necessarily need to include the reported date and time stored in the sleep duration prediction history DB 112. Furthermore, any of the learning algorithms described above may be used for machine learning. The instruction learning unit 104 then stores the generated learning model in the storage unit 110 as an instruction learning model 117 (second learning model).

[0046] As described above, the information processing device 10 according to this embodiment uses two learning models: the sleep time learning model 116 generated by the instruction learning unit 104 and the instruction learning model 117 generated by the sleep time prediction unit 105. Normally, using one learning model is desirable because it simplifies the configuration. However, a configuration in which one learning model is generated and used by machine learning requires a large amount of memory to be installed in the information processing device, and a long learning time is required. It may also be considered sufficient to prepare one learning model by machine learning in advance. However, since the characteristics of each facility in a nursing care or medical setting differ, it has become clear that it is difficult to deploy a single learning model in common to all facilities. Therefore, by dividing the learning model into two and generating and using each learning model in a compact form as described above, it becomes possible to generate and use these learning models for each facility according to the actual conditions of each facility.

[0047] Furthermore, prior art machine learning systems have been developed that directly predict events such as falls, aspiration, and incontinence based on recorded data from nursing and medical journals, but the accuracy of these predictions has been found to be unsatisfactory. Here, the events to be predicted for the subject are falls, aspiration, incontinence, etc., as described above, and experiments have shown that these events are highly correlated with sleep deprivation. For example, sleep deprivation can increase the likelihood of falls, cause distraction and lead to more eating problems, or impair decision-making, leading to missed urination timing and incontinence. Therefore, in the information processing device 10 according to this embodiment, the recorded data is included in the learning data to generate a sleep time learning model 116 that predicts sleep time using the sleep time learning unit 103, and the predicted sleep time is included in the learning data to generate an instruction learning model 117 that predicts warnings for falls, aspiration, incontinence, etc. using the instruction learning unit 104. In this way, by generating the instruction learning model 117 using sleep time, which is highly correlated with falls, aspiration, incontinence, etc., the accuracy of warning predictions can be improved.

[0048] As described above, the instruction learning unit 104 uses the morphological analysis unit 102 to break down the recorded data into morphemes through morphological analysis processing, and generates the instruction learning model 117 through machine learning using words and phrases related to troubles such as falls, aspiration, and incontinence, and predicted sleep duration. This makes it possible to improve the accuracy of predictions of warnings about falls, aspiration, incontinence, etc.

[0049] The sleep time prediction unit 105 is a functional unit that predicts the sleep time of a subject corresponding to recorded data registered in the recorded data DB 111, from the recorded data, using the sleep time learning model 116 stored in the storage unit 110. In this case, the sleep time prediction unit 105 uses the results of decomposition by the morphological analysis unit 102 for text data included in the recorded data. The sleep time prediction unit 105 stores the predicted sleep time in the sleep time prediction history DB 112 of the storage unit 110.

[0050] The instruction prediction unit 106 is a functional unit that predicts warnings of a subject's fall, aspiration, incontinence, etc., based on the record data registered in the record data DB 111 and the report date and time and predicted sleep time of the subject corresponding to the record data stored in the sleep time prediction history DB 112, using the instruction learning model 117 stored in the storage unit 110. In this case, the instruction prediction unit 106 uses the results of decomposition by the morphological analysis unit 102 for the text data included in the record data. Note that the report date and time stored in the sleep time prediction history DB 112 is not necessarily required for prediction by the instruction prediction unit 106.

[0051] The email generation unit 107 is a functional unit that generates an email message for notifying a warning predicted by the instruction prediction unit 106, using an email message DB 113 in the storage unit 110, which will be described later.

[0052] The notification unit 108 is a functional unit that notifies the warning predicted by the instruction prediction unit 106 by sending an email containing the email text generated by the email generation unit 107 via the communication unit 101 to the email address of the employee stored in the email destination DB 114 of the memory unit 110 described later.

[0053] 4, the storage unit 110 is a functional unit that stores a recorded data DB 111, a sleep duration prediction history DB 112, an email text DB 113, an email destination DB 114, a learning data DB 115, a sleep duration learning model 116, and an instruction learning model 117. The storage unit 110 is realized by the auxiliary storage device 705 shown in FIG.

[0054] The record data DB111 is a database that stores record data that is input and registered by a staff member at the staff member terminal 20. For example, record data such as those shown in Fig. 5(a) to Fig. 5(c) is registered in the record data DB111. As shown in Fig. 5(a) to Fig. 5(c), the record data includes an inputter indicating the staff member, the date and time when the record data was input, a care recipient indicating the subject, and report content that describes the subject's behavior, condition, etc.

[0055] The sleep time prediction history DB 112 is a database that stores the sleep time predicted by the sleep time prediction unit 105. For example, as shown in Fig. 9 described later, the sleep time prediction history DB 112 stores the date and time (report date and time) when the sleep time was predicted by the sleep time prediction unit 105, a care recipient indicating the subject, a care recipient ID that is identification information of the subject, and the predicted sleep time (predicted sleep time) in association with each other.

[0056] The email text DB 113 is a database that stores drafts of email text for notifying workers of warnings about subjects predicted by the instruction prediction unit 106.

[0057] The mail destination DB 114 is a database that stores the mail addresses of employees to whom the mail containing the mail text generated by the mail generating unit 107 is to be sent.

[0058] The learning data DB 115 is a database that stores learning data used in machine learning by the sleep time learning unit 103 and the instruction learning unit 104. The configuration of the learning data is as described above.

[0059] The sleep time learning model 116 is a learning model generated by the sleep time learning unit 103 as described above.

[0060] The instruction learning model 117 is a learning model generated by the instruction learning unit 104 as described above.

[0061] The above-mentioned communication unit 101, morphological analysis unit 102, sleep time learning unit 103, instruction learning unit 104, sleep time prediction unit 105, instruction prediction unit 106, email generation unit 107, and notification unit 108 are realized by, for example, executing a program by CPU 701 shown in Fig. 2. Note that at least some of the communication unit 101, morphological analysis unit 102, sleep time learning unit 103, instruction learning unit 104, sleep time prediction unit 105, instruction prediction unit 106, email generation unit 107, and notification unit 108 may be realized by a hardware circuit such as an ASIC (Application Specific Integrated Circuit).

[0062] Note that the functional units of the information processing device 10 shown in Fig. 4 conceptually illustrate their functions, and are not limited to such a configuration. For example, the multiple functional units illustrated as independent functional units in the information processing device 10 shown in Fig. 4 may be configured as a single functional unit. On the other hand, the function of a single functional unit in the information processing device 10 shown in Fig. 4 may be divided into multiple units and configured as multiple functional units. Furthermore, the functional units of the information processing device 10 do not need to be configured as distinct software modules as shown in Fig. 4, and it is sufficient that the functions of the functional units as a whole are realized by executing a program on the information processing device 10.

[0063] (Sleep time learning process flow of information processing device) 6 is a flowchart showing an example of the flow of the sleep time learning process of the information processing device according to Embodiment 1. The flow of the sleep time learning process of the information processing device 10 according to this embodiment will be described with reference to FIG.

[0064] <Step S11> The morphological analysis unit 102 of the information processing device 10 acquires the learning data stored in the learning data DB 115 to be used by the sleep time learning unit 103. The learning data here is as described above. Note that it may be the sleep time learning unit 103 that acquires the learning data from the learning data DB 115. Then, the process proceeds to step S12.

[0065] <Step S12> The morphological analysis unit 102 performs morphological analysis on the text data such as report contents in the record data included in the acquired learning data, and breaks it down into minimum units (morphemes) such as words, etc. Then, the process proceeds to step S13.

[0066] <Step S13> The sleep time learning unit 103 of the information processing device 10 generates a learning model for predicting sleep time by machine learning using the result of the morphological analysis process performed by the morphological analysis unit 102 on the learning data stored in the learning data DB 115. Then, the process proceeds to step S14.

[0067] <Step S14> The sleep time learning unit 103 stores the generated learned learning model in the storage unit 110 as the sleep time learning model 116. Then, the sleep time learning process ends.

[0068] (Flow of warning learning process of information processing device) 7 is a flowchart showing an example of the flow of the warning learning process of the information processing device according to the first embodiment. The flow of the warning learning process of the information processing device 10 according to this embodiment will be described with reference to FIG.

[0069] <Step S21> The morphological analysis unit 102 of the information processing device 10 acquires the learning data stored in the learning data DB 115 to be used by the instruction learning unit 104. The learning data here is as described above. Note that the instruction learning unit 104 may acquire the learning data from the learning data DB 115. Then, the process proceeds to step S22.

[0070] <Step S22> The morphological analysis unit 102 performs morphological analysis on the text data such as report contents in the record data included in the acquired learning data, and breaks it down into minimum units (morphemes) such as words, etc. Then, the process proceeds to step S23.

[0071] <Step S23> The instruction learning unit 104 of the information processing device 10 generates a learning model for predicting warnings about a fall, aspiration, incontinence, etc. for a subject by machine learning using the results of the morphological analysis process performed by the morphological analysis unit 102 on the learning data stored in the learning data DB 115. Then, the process proceeds to step S24.

[0072] <Step S24> The instruction learning unit 104 stores the generated learned learning model in the storage unit 110 as the instruction learning model 117. Then, the warning learning process ends.

[0073] (Sleep time prediction process flow of information processing device) Fig. 8 is a flowchart showing an example of the flow of sleep time prediction processing of the information processing device according to the first embodiment. Fig. 9 is a diagram showing an example of the sleep time prediction history of the information processing device according to the first embodiment. The flow of sleep time prediction processing of the information processing device 10 according to this embodiment will be described with reference to Figs. 8 and 9.

[0074] <Step S31> The morphological analysis unit 102 of the information processing device 10 acquires the record data of the subject registered by the staff terminal 20 from the record data DB 111. Note that the sleep time prediction unit 105 may acquire the record data from the record data DB 111. Then, the process proceeds to step S32.

[0075] <Step S32> The morphological analysis unit 102 breaks down text data such as report contents in the acquired record data into minimum units (morphemes) such as words through morphological analysis processing, and then the process proceeds to step S33.

[0076] <Step S33> The sleep time prediction unit 105 of the information processing device 10 predicts the sleep time of the subject corresponding to the record data registered in the record data DB 111, using the sleep time learning model 116 stored in the storage unit 110. In this case, the sleep time prediction unit 105 uses the results of decomposition by the morphological analysis unit 102 for the text data included in the record data. Then, the process proceeds to step S34.

[0077] <Step S34> The sleeping time prediction unit 105 stores the predicted sleeping time in the sleeping time prediction history DB 112 of the storage unit 110 as shown in Fig. 9. That is, the sleeping time prediction unit 105 stores, in the sleeping time prediction history DB 112, for example, the date and time when the sleeping time was predicted by the sleeping time prediction unit 105 (reported date and time), the care recipient indicating the subject, the care recipient ID which is the identification information of the subject, and the predicted sleeping time (predicted sleeping time) in association with each other. Then, the sleeping time prediction process ends.

[0078] (Flow of warning prediction process of information processing device) Fig. 10 is a flowchart showing an example of the flow of warning prediction processing of the information processing device according to the first embodiment. Fig. 11 is a diagram showing an example of the text of an email displayed on the employee terminal according to the first embodiment. The flow of warning prediction processing of the information processing device 10 according to this embodiment will be described with reference to Figs. 10 and 11.

[0079] <Step S41> The morphological analysis unit 102 of the information processing device 10 acquires the subject's predicted sleeping time (sleep time data) and the report date and time stored in the sleep time prediction history DB 112. Note that the instruction prediction unit 106 may acquire the predicted sleeping time from the sleep time prediction history DB 112. Then, the process proceeds to step S42.

[0080] <Step S42> The morphological analysis unit 102 acquires the record data used to predict the predicted sleep time acquired in step S41 from the record data registered in the record data DB 111. Note that the instruction prediction unit 106 may acquire the record data from the record data DB 111. Then, the process proceeds to step S43.

[0081] <Step S43> The morphological analysis unit 102 breaks down text data such as report contents in the acquired record data into minimum units (morphemes) such as words through morphological analysis processing, and then the process proceeds to step S44.

[0082] <Step S44> The instruction prediction unit 106 of the information processing device 10 uses the instruction learning model 117 stored in the storage unit 110 to predict warnings for the subject's fall, aspiration, incontinence, etc., based on the acquired predicted sleep time and report date and time of the subject, as well as the recorded data. In this case, the instruction prediction unit 106 uses the results of decomposition by the morphological analysis unit 102 for the text data included in the recorded data. Then, the process proceeds to step S45.

[0083] <Step S45> The email generation unit 107 of the information processing device 10 uses the email text DB 113 of the storage unit 110 to generate an email message for notifying the worker of the warning predicted by the instruction prediction unit 106. Then, the notification unit 108 of the information processing device 10 notifies the worker of the warning predicted by the instruction prediction unit 106 by sending an email including the email message generated by the email generation unit 107 to the worker's email address stored in the email destination DB 114 of the storage unit 110 via the communication unit 101. Then, as shown in FIG. 11 , the email indicating the warning notified by the notification unit 108 is displayed on the worker terminal 20, allowing the worker to confirm the warning. Then, the warning prediction process ends.

[0084] It should be noted that the processes included in the flowcharts of FIGS. 6, 7, 8, and 10 are merely examples and do not represent limiting aspects.

[0085] As described above, the information processing device 10 according to this embodiment includes the sleep time learning model 116, which is machine-learned to predict the sleeping time of a subject based on record data about the subject being observed by a medical or nursing care worker, and the instruction learning model 117, which is machine-learned to predict a warning as an instruction for the subject based on the sleeping time of the subject predicted using the sleep time learning model 116. In this way, by providing two learning models, the learning time for the learning model can be reduced, while an appropriate warning can be predicted by predicting the sleeping time of the subject being observed.

[0086] [Second embodiment] The information processing system 1 according to the second embodiment will be described, focusing on the differences from the information processing system 1 according to the first embodiment. In the first embodiment, the operation of predicting a warning regarding a subject, which is an example of an instruction, using the instruction learning model 117 was described. In this embodiment, the operation of predicting a menu for a subject, which is an example of an instruction, using the instruction learning model 117 will be described. Note that the overall configuration of the information processing system 1 according to this embodiment, the hardware configuration of the information processing device 10 and the employee terminal 20, and the configuration of the functional blocks of the information processing device 10 are the same as the configurations described in the first embodiment. In addition, the sleep time learning process and sleep time prediction process of the information processing device 10 according to this embodiment are also the same as the processes described in the first embodiment.

[0087] (Configuration and operation of functional blocks of information processing device) 12 is a diagram showing an example of recorded data in the information processing system according to the second embodiment. As described above, the configuration of the information processing device 10 according to this embodiment is the same as that of the information processing device 10 according to the first embodiment, but different operations will be described below.

[0088] The communication unit 101 performs data communication with the staff terminal 20 and the like via the network I / F 709. For example, the communication unit 101 receives record data as shown in FIG. 12 input by a staff member from the staff terminal 20 via the network N, and registers the record data in the record data DB 111 of the storage unit 110. The example of the record data shown in FIG. 12 includes an inputter indicating the staff member, the date and time when the record data was input, a care recipient indicating the subject, and report content indicating the content and situation of food ingested by the subject. Note that the record data is not limited to that shown in FIG. 12, and may include record data of the content shown in the first embodiment described above.

[0089] The instruction learning unit 104 is a functional unit that generates a learning model for predicting a food menu (an example of an instruction) to be provided to a subject through machine learning using the learning data stored in the learning data DB 115 of the storage unit 110. Here, the menu refers to, for example, a menu for three meals (breakfast, lunch, and dinner) such as curry and stew, or a snack (snack) such as pudding or ice cream. Specifically, the instruction learning unit 104 generates the learning model through machine learning using the results of morphological analysis performed by the morphological analysis unit 102 on the learning data stored in the learning data DB 115. Here, the learning data stored in the learning data DB 115 used by the instruction learning unit 104 is, for example, data in which the recorded data registered in the recorded data DB 111 and the reported date and time and predicted sleep time predicted by the sleep time prediction unit 105 and stored in the sleep time prediction history DB 112 of the storage unit 110 are associated as labels with the menu provided to the subject as optimal or suitable after the recorded data was registered by the staff member. The machine learning is as described above. Then, the instruction learning unit 104 stores the generated learning model in the storage unit 110 as the instruction learning model 117.

[0090] Morphological analysis can be used to break down recorded data, including the content and circumstances of the food consumed by the subject, into morphemes, and then train the system based on meal-related terms and dates and times to understand each subject's food intake and preferences. While eating everything is desirable in caregiving or medical settings, forcing individuals to eat their meals is not an option. If a person consistently fails to finish any of their meals (breakfast, lunch, or dinner), a condition known as a reduced appetite, modifying the menu to suit the individual's preferences can increase the likelihood of the individual eating. Furthermore, some caregiving or medical facilities prepare large quantities of food, making it difficult to easily change the menu for all three meals (breakfast, lunch, and dinner). Even in such cases, individual preferences can be tailored to snacks such as pudding, yogurt, ice cream, sweet breads, and sandwiches. Increasing the amount of food eaten at each meal (breakfast, lunch, and dinner) or between meals (snacks) can also help maintain calorie intake.

[0091] Experiments have also shown a strong correlation between lack of sleep and aspiration. In such cases, it has been found that in order to reduce the risk of aspiration, it is preferable to change snacks to items that are less likely to cause choking, such as pudding, yogurt, or ice cream. This also reduces the risk of choking. In contrast, it is not recommended to provide highly elastic jellies, etc., as they pose a very high risk of choking due to aspiration. While highly elastic jellies are not often provided in nursing care or medical facilities, they are sometimes brought in as souvenirs by the patients' families. It is preferable to change the snack menu to avoid giving these souvenirs to patients.

[0092] The instruction prediction unit 106 uses the instruction learning model 117 stored in the storage unit 110 to predict a food menu to be served to a subject based on the record data registered in the record data DB 111 and the reported date and time and predicted sleep time of the subject corresponding to the record data stored in the sleep time prediction history DB 112. In this case, the instruction prediction unit 106 uses the results of decomposition by the morphological analysis unit 102 for the text data included in the record data. Note that the reported date and time stored in the sleep time prediction history DB 112 is not necessarily required for prediction by the instruction prediction unit 106.

[0093] The email generation unit 107 uses the email text DB 113 in the storage unit 110 to generate an email text for notifying the menu predicted by the instruction prediction unit 106.

[0094] The notification unit 108 is a functional unit that notifies the menu predicted by the instruction prediction unit 106 by sending an email containing the email text generated by the email generation unit 107 to the email address of the employee stored in the email destination DB 114 of the memory unit 110 via the communication unit 101.

[0095] (Flow of menu learning process of information processing device) 13 is a flowchart showing an example of the flow of menu learning processing by the information processing device according to embodiment 2. The flow of menu learning processing by the information processing device 10 according to this embodiment will be described with reference to FIG.

[0096] <Step S51> The morphological analysis unit 102 of the information processing device 10 acquires the learning data stored in the learning data DB 115 to be used by the instruction learning unit 104. The learning data here is as described above. Note that the instruction learning unit 104 may acquire the learning data from the learning data DB 115. Then, the process proceeds to step S52.

[0097] <Step S52> The morphological analysis unit 102 performs morphological analysis on the text data such as report contents in the record data included in the acquired learning data, and breaks it down into minimum units (morphemes) such as words, etc. Then, the process proceeds to step S53.

[0098] <Step S53> The instruction learning unit 104 of the information processing device 10 generates a learning model for predicting a food menu to be served to a subject by machine learning using the results of the morphological analysis process performed by the morphological analysis unit 102 on the learning data stored in the learning data DB 115. Then, the process proceeds to step S54.

[0099] <Step S54> The instruction learning unit 104 stores the generated trained learning model in the storage unit 110 as the instruction learning model 117. Then, the menu learning process ends.

[0100] (Flow of menu prediction processing by information processing device) 14 is a flowchart showing an example of the flow of menu prediction processing by the information processing device according to embodiment 2. The flow of menu prediction processing by the information processing device 10 according to this embodiment will be described with reference to FIG.

[0101] <Step S61> The morphological analysis unit 102 of the information processing device 10 acquires the subject's predicted sleeping time (sleep time data) and the report date and time stored in the sleep time prediction history DB 112. Note that the instruction prediction unit 106 may acquire the predicted sleeping time from the sleep time prediction history DB 112. Then, the process proceeds to step S62.

[0102] <Step S62> The morphological analysis unit 102 acquires the record data used to predict the predicted sleep time acquired in step S61 from the record data registered in the record data DB 111. Note that the instruction prediction unit 106 may acquire the record data from the record data DB 111. Then, the process proceeds to step S63.

[0103] <Step S63> The morphological analysis unit 102 breaks down text data such as report contents in the acquired record data into minimum units (morphemes) such as words through morphological analysis processing, and then the process proceeds to step S64.

[0104] <Step S64> The instruction prediction unit 106 of the information processing device 10 predicts a food menu to be served to the subject from the acquired predicted sleep time and reported date and time of the subject, as well as the recorded data, using the instruction learning model 117 stored in the storage unit 110. In this case, the instruction prediction unit 106 uses the results of decomposition by the morphological analysis unit 102 for the text data included in the recorded data. Then, the process proceeds to step S65.

[0105] <Step S65> The email generation unit 107 of the information processing device 10 uses the email text DB 113 in the storage unit 110 to generate an email message for notifying the staff member of the menu predicted by the instruction prediction unit 106. The notification unit 108 of the information processing device 10 then notifies the staff member of the menu predicted by the instruction prediction unit 106 by sending an email containing the email message generated by the email generation unit 107 via the communication unit 101 to the staff member's email address stored in the email destination DB 114 in the storage unit 110. The staff member can then confirm the menu by seeing the email indicating the menu notified by the notification unit 108 displayed on the staff member terminal 20, as in the case shown in FIG. 11 above. The menu prediction process then ends.

[0106] It should be noted that the processes included in the flowcharts of FIGS. 13 and 14 are merely examples and do not represent limiting aspects.

[0107] As described above, in the information processing device 10 according to this embodiment, the instruction learning model 117 is a learning model that has been machine-learned to predict a food menu to be provided to a subject as an instruction to the subject, based on the subject's sleeping time predicted using the sleep time learning model 116. This allows for two learning models to be used, thereby reducing the learning time of the learning model, and predicting an appropriate menu by predicting the sleeping time of the subject to be observed.

[0108] Each function of the above-described embodiments can be realized by one or more processing circuits. Here, the term "processing circuit" includes a processor programmed to execute each function by software, such as a processor implemented by an electronic circuit, and devices designed to execute each of the above-described functions, such as an ASIC, a DSP (Digital Signal Processor), an FPGA (Field-Programmable Gate Array), an SoC (System on a Chip), a GPU (Graphics Processing Unit), or a conventional circuit module.

[0109] In each of the above-described embodiments, when at least one of the functional units of the information processing device 10 is realized by executing a program, the program is provided by being pre-installed in a ROM or the like. In each of the above-described embodiments, the program executed by the information processing device 10 may be provided by being recorded in an installable or executable file format on a computer-readable recording medium such as a CD-ROM (Compact Disc Read Only Memory), a flexible disk (FD), a CD-R (Compact Disk-Recordable), or a DVD (Digital Versatile Disc). In each of the above-described embodiments, the program executed by the information processing device 10 may be stored on a computer connected to a network such as the Internet and provided by being downloaded via the network. In each of the above-described embodiments, the program executed by the information processing device 10 may be provided or distributed via a network such as the Internet. In each of the above-described embodiments, the program executed by the information processing device 10 has a modular configuration including at least one of the above-described functional units. In terms of actual hardware, the CPU 701 reads and executes the program from the above-described auxiliary storage device 705, thereby loading and generating the above-described functional units into a main storage device (e.g., RAM 703).

[0110] The aspects of the present invention are as follows. <1> A first learning model that is machine-learned to predict the sleep time of a subject based on recorded data on the subject that is observed by a medical or nursing care worker, and a second learning model that is machine-learned to predict instructions for the subject based on the subject's sleep time predicted using the first learning model; The information processing device is provided with: <2> a first prediction unit that predicts the sleep duration of the subject using the first learning model from record data about the subject created by the worker; a second prediction unit that predicts instructions to the subject using the second learning model from the recorded data used for prediction by the first prediction unit and the sleeping time of the subject predicted by the first prediction unit; The said further comprising <1> The information processing device is described in <3> The apparatus further includes a notification unit that notifies the worker of the instruction predicted by the second prediction unit. <2> The information processing device is described in <4> a first learning unit that generates the first learning model for predicting the subject's sleep time by machine learning using first learning data including record data about the subject; a second learning unit that generates the second learning model for predicting instructions to the subject by machine learning using second learning data including record data on the subject and the subject's sleeping time predicted by the first learning model; and The said further comprising <1> ~ <3> 1 is an information processing device according to any one of the preceding claims. <5> The first learning model is a learning model that is machine-learned based on recorded data that represents the behavior or state of the subject. <1> ~ <4> 1 is an information processing device according to any one of the preceding claims. <6> the second learning model is a learning model trained by machine learning to predict a warning about the subject's behavior as the instruction to the subject, based on the subject's sleep time predicted using the first learning model; <1> ~ <5> 1 is an information processing device according to any one of the preceding claims. <7> The second learning model is a learning model that is machine-learned to predict a menu of foods to be provided to the subject as the instruction to the subject, based on the sleeping time of the subject predicted using the first learning model. <1> ~ <5> 1 is an information processing device according to any one of the preceding claims. <8> The menu is a menu for three meals. <7> The information processing device is described in <9> The menu is a snack menu. <7> The information processing device is described in <10> a first prediction step of predicting the sleep time of a subject from record data about the subject created by a medical or care worker using a first learning model that has been machine-learned to predict the sleep time of the subject based on record data about the subject that the worker is observing; a second prediction step of predicting instructions to the subject from the recorded data used for the prediction in the first prediction step and the subject's sleeping time predicted in the first prediction step, using a second learning model that has been machine-learned to predict instructions to the subject based on the subject's sleeping time predicted using the first learning model; The information processing method has the following features. <11> On the computer, a first prediction step of predicting the sleep time of a subject from record data about the subject created by a medical or care worker using a first learning model that has been machine-learned to predict the sleep time of the subject based on record data about the subject that the worker is observing; a second prediction step of predicting instructions to the subject from the recorded data used for the prediction in the first prediction step and the subject's sleeping time predicted in the first prediction step, using a second learning model that has been machine-learned to predict instructions to the subject based on the subject's sleeping time predicted using the first learning model; This is a program for executing the above. [Explanation of symbols]

[0111] 1. Information Processing Systems 10. Information processing equipment 20, 20a, 20b Staff terminal 101 Communications Department 102 Morphological analysis section 103 Sleep Time Learning Department 104 Instruction Learning Section 105 Sleep Time Prediction Unit 106 Instruction Prediction Unit 107 Email Generation Unit 108 Notification Department 110 Storage section 111 Recorded Data DB 112 Sleep time prediction history DB 113 Email Text DB 114 Mailing Address DB 115 Learning Data DB 116 Sleep Time Learning Model 117 Instructional Learning Model 701 CPU 702 ROM 703 RAM 705 Auxiliary storage 706 Recording Media 707 Media Drive 708 Display 709 Network I / F 710 Bus 711 keyboard 712 Mouse 713 DVD 714 DVD drive 801 CPU 802 ROM 803 RAM 804 EEPROM 805 Imaging unit 806 Imaging I / F 807 Acceleration and direction sensor 808 GNSS receiver 809 Bus Line 810 Telecommunications Circuit 810a antenna 811 Near field communication circuit 811a antenna 812 Mike 813 Speaker 814 Sound input / output I / F 815 Display 816 External device connection I / F 817 Vibrator 818 Touch Panel N Network [Prior art documents] [Patent documents]

[0112] [Patent Document 1] Japanese Patent Publication No. 2022-169193 [Patent Document 2] Japanese Patent Application Publication No. 2019-017499 [Patent Document 3] Patent Publication No. 2021-064327 [Patent Document 4] Patent No. 7095598 [Patent Document 5] Japanese Patent Application Laid-Open No. 2003-122852

Claims

1. A first learning model that is machine-learned to predict the sleep time of a subject based on recorded data on the subject that is observed by a medical or nursing care worker, the first learning model being created by machine learning to predict the sleep time of the subject; a second learning model that is machine-learned to predict instructions for the subject based on the subject's sleep time predicted using the first learning model; An information processing device comprising:

2. a first prediction unit that predicts the sleep duration of the subject using the first learning model from record data on the subject created by the worker; a second prediction unit that predicts instructions to the subject using the second learning model from the recorded data used for prediction by the first prediction unit and the sleeping time of the subject predicted by the first prediction unit; The information processing device according to claim 1 , further comprising:

3. The information processing apparatus according to claim 2 , further comprising a notification unit that notifies the worker of the instruction predicted by the second prediction unit.

4. a first learning unit that generates the first learning model for predicting the sleep time of the subject by machine learning using first learning data including record data on the subject; a second learning unit that generates the second learning model for predicting instructions to the subject by machine learning using second learning data including record data on the subject and the subject's sleeping time predicted by the first learning model; and The information processing device according to claim 1 or 2, further comprising:

5. The information processing device according to claim 1 , wherein the first learning model is a learning model that is machine-learned based on recorded data that represents the behavior or state of the subject.

6. The information processing device described in claim 1 or 2, wherein the second learning model is a learning model machine-learned to predict a warning regarding the subject's behavior as the instruction to the subject based on the subject's sleep time predicted using the first learning model.

7. 3. The information processing device according to claim 1, wherein the second learning model is a learning model trained by machine learning to predict a food menu to be provided to the subject as the instruction to the subject based on the subject's sleep time predicted using the first learning model.

8. The information processing device according to claim 7 , wherein the menu is a menu for three meals.

9. The information processing device according to claim 7 , wherein the menu is a menu for snacks.

10. a first prediction step of predicting the sleep time of a subject from record data about the subject created by a medical or care worker using a first learning model that has been machine-learned to predict the sleep time of the subject based on record data about the subject that the worker is observing; a second prediction step of predicting instructions to the subject from the recorded data used for the prediction in the first prediction step and the subject's sleeping time predicted in the first prediction step, using a second learning model that has been machine-learned to predict instructions to the subject based on the sleeping time of the subject predicted using the first learning model; An information processing method comprising:

11. On the computer, a first prediction step of predicting the sleep time of a subject from record data about the subject created by a medical or care worker using a first learning model that has been machine-learned to predict the sleep time of the subject based on record data about the subject that the worker is observing; a second prediction step of predicting instructions to the subject from the recorded data used for the prediction in the first prediction step and the subject's sleeping time predicted in the first prediction step, using a second learning model that has been machine-learned to predict instructions to the subject based on the sleeping time of the subject predicted using the first learning model; A program to execute.

Images