FIM value estimation device, FIM value estimation method, FIM value estimation program, artificial intelligence learning device, artificial intelligence learning method, and artificial intelligence learning program

Abstract

This makes it possible to simply obtain the FIM score of the subject without relying on sleep data. [Solution] The FIM value estimation device according to the present disclosure includes an input data receiving unit, an estimation unit, and an estimated data output unit. The input data receiving unit receives input data including a simplified FIM value, which represents the FIM value of the person being evaluated using a plurality of values ​​less than seven for fewer than 18 items. The estimation unit inputs the input data received by the input data receiving unit into a trained artificial intelligence, causing the artificial intelligence to calculate estimated data, which represents the FIM value of the person being evaluated using seven values ​​for the 18 items. The estimated data output unit outputs the estimated data calculated by the artificial intelligence.

Description

[Technical Field]

[0001] The present invention relates to an FIM value estimation device, an FIM value estimation method, an FIM value estimation program, an artificial intelligence learning device, an artificial intelligence learning method, and an artificial intelligence learning program, which estimate the FIM value (evaluation value according to the Functional Independence Assessment Method) of a subject to be evaluated. [Background technology]

[0002] The Functional Independence Measure (FIM), developed in the United States, is known as a method for assessing the degree to which patients and care recipients are able to independently perform activities of daily living (ADL). Because the FIM can be used to assess activities of daily living, particularly the level of caregiver burden, it is a well-known method widely used in Japan, including in the field of rehabilitation. The FIM is also an effective tool for occupational therapy evaluation in occupational therapy.

[0003] The FIM consists of 18 motor and cognitive items, each of which is evaluated on a 7-point scale ranging from 1 to 7 points. The cognitive items are five items: understanding, expression, social interaction, problem solving, and memory. The motor items are 13 items in total, including eating, grooming, wiping, dressing, toileting, urinary management, bowel management, transfers (activities of getting up and down), and mobility. The FIM evaluation items are thus broad, and the evaluation values ​​are subdivided. For this reason, determining the FIM evaluation value (abbreviated as "FIM value" in this application) is a time-consuming task.

[0004] As one means for solving this problem, the present inventor has completed an invention that uses artificial intelligence to estimate FIM scores based on the sleep data and basic data of the person being evaluated (see Patent Document 1). The sleep data is data related to the subject's sleep, such as breathing, pulse, sleep duration, and sleep rhythm, and is acquired using a sleep sensor. The basic data is data related to the subject's body, such as the subject's age, sex, height, and weight, and is acquired through a doctor's diagnosis, etc.

[0005] The inventors of the present application have been searching for a way to simply obtain FIM scores without relying on sleep data. While Patent Documents 2 and 3 disclose techniques for estimating FIM scores at a later point in time from values ​​at an earlier point in time, they, like Patent Document 1, do not disclose a technique for estimating a wide range of FIM scores from limited FIM values. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Patent No. 6994262 [Patent Document 2] WO-A-2021-075017 publication [Patent Document 3] WO-A-2021-033281 publication Summary of the Invention [Problem to be solved by the invention]

[0007] The present invention was achieved through the inventor's research as described above, and aims to provide an FIM value estimation device, an FIM value estimation method, and an FIM value estimation program that enable the FIM value of a subject to be evaluated to be easily obtained without relying on sleep data. Another aim of the present invention is to provide an artificial intelligence learning device, an artificial intelligence learning method, and an artificial intelligence learning program that train an artificial intelligence that can use the FIM value estimation device, FIM value estimation method, and FIM value estimation program. [Means for solving the problem]

[0008] To achieve the above object, a first aspect of the present invention provides an FIM value estimation device comprising an input data receiving unit, an estimation unit, and an estimated data output unit. The input data receiving unit receives input data including a simplified FIM value, which represents the FIM value of an individual being evaluated using a plurality of values ​​less than seven for fewer than 18 items. The estimation unit inputs the input data received by the input data receiving unit into a trained artificial intelligence, causing the artificial intelligence to calculate estimated data, which represents the FIM value of the individual being evaluated using seven values ​​for each of the 18 items. The estimated data output unit outputs the estimated data calculated by the artificial intelligence.

[0009] According to this configuration, by preparing a simplified FIM score for the subject, an FIM score expressed by seven values ​​for 18 items can be obtained. In other words, the subject's FIM score can be simply obtained without relying on sleep data. The trained artificial intelligence may be part of the FIM score estimation device of this configuration, or it may be an external device, such as one placed on an external cloud server.

[0010] A second aspect of the present invention is an FIM value estimation device according to the first aspect, in which the simplified FIM value includes at least one item out of 13 items belonging to motor items and at least one item out of 5 items belonging to cognitive items. According to this configuration, the FIM value can be obtained with higher accuracy.

[0011] A third aspect of the present invention is an FIM value estimation device according to the second aspect, in which the simplified FIM value includes two motor items, namely, bed transfer and walking, and two cognitive items, namely, problem-solving ability and memory, and each item is represented by three values. According to this configuration, the FIM value can be obtained with even higher accuracy.

[0012] A fourth aspect of the present invention is an FIM value estimation device according to any one of the first to third aspects, wherein the input data includes basic data of the subject, including at least one item selected from the subject's age, BMI (body mass index), height, and weight. According to this configuration, the FIM value can be obtained with higher accuracy. A fifth aspect of the present invention is the FIM value estimation device according to any one of the first to fourth aspects, wherein the input data includes living environment data of the subject, and the living environment data is expressed by a numerical value. According to this configuration, the FIM value can be obtained with higher accuracy.

[0013] A sixth aspect of the present invention is a method for estimating an FIM value, comprising an input data reception process, an estimation process, and an estimated data output process. The input data reception process receives input data including a simplified FIM value, which represents the FIM value of a subject to be evaluated using a plurality of values ​​less than seven for fewer than 18 items. The estimation process inputs the input data received by the input data reception process into a trained artificial intelligence, causing the artificial intelligence to calculate estimated data, which represents the FIM value of the subject to be evaluated using seven values ​​for each of the 18 items. The estimated data output process outputs the estimated data calculated by the artificial intelligence.

[0014] According to this configuration, by preparing a simplified FIM score for the subject, it is possible to obtain an FIM score expressed by seven values ​​for 18 items. In other words, the FIM score for the subject can be obtained simply without relying on sleep data.

[0015] A seventh aspect of the present invention is the FIM value estimation method according to the sixth aspect, in which the simplified FIM value includes at least one item out of 13 items belonging to motor items and at least one item out of 5 items belonging to cognitive items. According to this configuration, the FIM value can be obtained with higher accuracy.

[0016] An eighth aspect of the present invention is the FIM score estimation method according to the seventh aspect, in which the simplified FIM score includes two motor items, namely, bed transfer and walking, and two cognitive items, namely, problem-solving ability and memory, and each item is represented by three values. According to this configuration, the FIM value can be obtained with even higher accuracy.

[0017] A ninth aspect of the present invention is a method for estimating an FIM value according to any one of the sixth to eighth aspects, wherein the input data includes basic data of the subject, including at least one item selected from the subject's age, BMI (body mass index), height, and weight. According to this configuration, the FIM value can be obtained with higher accuracy. A tenth aspect of the present invention is the FIM score estimation method according to any one of the sixth to ninth aspects, wherein the input data includes living environment data of the subject, and the living environment data is expressed by a numerical value. According to this configuration, the FIM value can be obtained with higher accuracy.

[0018] An eleventh aspect of the present invention is an FIM value estimation program that, when read by a computer, causes the computer to execute the FIM value estimation method according to any one of the sixth to tenth aspects. According to the program having this configuration, the FIM value estimation method according to any one of the sixth to tenth aspects is realized by a computer.

[0019] A twelfth aspect of the present invention is an artificial intelligence learning device comprising an input data accepting unit, a teacher data accepting unit, and a learning unit. The input data accepting unit accepts input data including simplified FIM values ​​in which the FIM scores of an individual to be evaluated are expressed using a plurality of values ​​less than seven for fewer than 18 items. The teacher data accepting unit accepts teacher data corresponding to the input data and in which the FIM scores of the individual to be evaluated are expressed using seven values ​​for the 18 items. The learning unit inputs the input data accepted by the input data accepting unit and the teacher data accepted by the teacher data accepting unit into an artificial intelligence, thereby training the artificial intelligence to estimate the teacher data from the input data.

[0020] According to this configuration, an artificial intelligence usable for the FIM value estimation device according to the first aspect is constructed by learning. Note that the artificial intelligence may be a part of the artificial intelligence learning device of this configuration, or may be an external device, such as one placed on an external cloud server.

[0021] A thirteenth aspect of the present invention is an artificial intelligence learning device according to the twelfth aspect, in which the simplified FIM value includes at least one item out of thirteen items belonging to motor items and at least one item out of five items belonging to cognitive items. According to this configuration, artificial intelligence usable for the FIM value estimation device according to the second aspect is constructed by learning.

[0022] A fourteenth aspect of the present invention is an artificial intelligence learning device according to the thirteenth aspect, in which the simplified FIM score includes two items belonging to the motor category, namely, bed transfer and walking, and two items belonging to the cognitive category, namely, problem-solving ability and memory, and each item is represented by three values. According to this configuration, artificial intelligence usable for the FIM value estimation device according to the third aspect is constructed by learning.

[0023] A fifteenth aspect of the present invention is an artificial intelligence learning device according to any one of the twelfth to fourteenth aspects, wherein the input data includes basic data of the person being evaluated, including at least one item selected from the age, BMI (body mass index), height, and weight of the person being evaluated. According to this configuration, artificial intelligence usable for the FIM value estimation device according to the fourth aspect is constructed by learning. According to a sixteenth aspect of the present invention, there is provided an artificial intelligence learning device according to any one of the twelfth to fifteenth aspects, wherein the input data includes living environment data of the subject to be evaluated, and the living environment data is expressed by a numerical value. According to this configuration, artificial intelligence usable for the FIM value estimation device according to the fifth aspect is constructed by learning.

[0024] A seventeenth aspect of the present invention is an artificial intelligence learning method comprising an input data receiving process, a teacher data receiving process, and a learning process. The input data receiving process receives input data including simplified FIM values ​​in which the FIM scores of an individual to be evaluated are expressed using fewer than seven values ​​for fewer than 18 items. The teacher data receiving process receives teacher data corresponding to the input data and in which the FIM scores of the individual to be evaluated are expressed using seven values ​​for the 18 items. The learning process inputs the input data received by the input data receiving unit and the teacher data received by the teacher data receiving unit into an artificial intelligence, thereby training the artificial intelligence to estimate the teacher data from the input data. According to this configuration, artificial intelligence that can be used in the FIM value estimation method according to the sixth aspect is constructed by learning.

[0025] An 18th aspect of the present invention is an artificial intelligence learning method according to the 17th aspect, in which the simplified FIM value includes at least one item out of 13 items belonging to motor items and at least one item out of 5 items belonging to cognitive items. According to this configuration, artificial intelligence that can be used in the FIM value estimation method according to the seventh aspect is constructed by learning.

[0026] A 19th aspect of the present invention is an artificial intelligence learning method according to the 18th aspect, in which the simplified FIM score includes two motor items, namely, bed transfer and walking, and two cognitive items, namely, problem-solving ability and memory, and each item is represented by three values. According to this configuration, artificial intelligence that can be used in the FIM value estimation method according to the eighth aspect is constructed by learning.

[0027] A twentieth aspect of the present invention is an artificial intelligence learning method according to any one of the seventeenth to nineteenth aspects, wherein the input data includes basic data of the subject to be evaluated, including at least one item selected from the age, BMI (body mass index), height, and weight of the subject to be evaluated. According to this configuration, artificial intelligence that can be used in the FIM value estimation method according to the ninth aspect is constructed by learning. A 21st aspect of the present invention is an artificial intelligence learning method according to any one of the 17th to 20th aspects, wherein the input data includes living environment data of the subject to be evaluated, and the living environment data is expressed by a numerical value. According to this configuration, artificial intelligence that can be used in the FIM value estimation method according to the tenth aspect is constructed by learning.

[0028] A 22nd aspect of the present invention is an artificial intelligence learning program which, when read by a computer, causes the computer to execute the artificial intelligence learning method according to any one of the 17th to 21st aspects. According to the program with this configuration, the artificial intelligence learning method according to any one of the seventeenth to twenty-first aspects is realized by a computer. [Effects of the Invention]

[0029] As described above, the present invention provides an FIM value estimation device, an FIM value estimation method, and an FIM value estimation program that enable the FIM value of a subject to be evaluated to be easily obtained without relying on sleep data. The present invention also provides an artificial intelligence learning device, an artificial intelligence learning method, and an artificial intelligence learning program that train an artificial intelligence that can use the FIM value estimation device, the FIM value estimation method, and the FIM value estimation program. [Brief explanation of the drawings]

[0030] [Figure 1] FIG. 1 is a schematic explanatory diagram illustrating an outline of data processing by an FIM value estimation device according to an embodiment of the present invention. [Figure 2] FIG. 2 is a block diagram illustrating the configuration of an FIM value estimation device that executes the data processing of FIG. 1. [Figure 3] 3 is a flowchart illustrating the flow of processing of an FIM value estimation method implemented by the FIM value estimation device illustrated in FIG. 2. [Figure 4] 3 is a flowchart illustrating the flow of processing of an artificial intelligence learning method realized by the FIM value estimation device illustrated in FIG. 2. [Figure 5] FIG. 3 is a schematic diagram illustrating the conceptual configuration of artificial intelligence used by the FIM value estimation device illustrated in FIG. 2. DETAILED DESCRIPTION OF THE INVENTION

[0031] Table 1 shows all 18 items of the FIM. The items related to motor function are 13 items: eating, grooming, wiping, dressing (upper), dressing (lower), toileting, bladder control, bowel control, transfer (to bed), transfer (toilet), transfer (to bathtub), mobility (walking), and mobility (stairs). The items related to cognitive function are five items: understanding, expression, social interaction, problem solving, and memory. The evaluation value of each item is expressed as a number on a seven-point scale.

[0032] [Table 1]

[0033] FIG. 1 is a schematic diagram illustrating an example of data processing performed by an FIM value estimation device according to an embodiment of the present invention. The FIM value estimation device operates roughly as follows: For an FIM value assessment subject, such as a patient or care recipient, evaluation values ​​expressed in three numerical levels are prepared for four FIM value items: bed transfer, walking mobility, problem-solving ability, and memory. In other words, a simplified FIM value is prepared for the original FIM value, which is expressed in seven numerical levels for a total of 18 items. This simplified FIM value can be obtained by an individual providing treatment or care to the assessment subject, such as an occupational therapist or caregiver. Because it is a simplified FIM value, its evaluation is easier than evaluating an FIM value with 18 items and seven values.

[0034] Furthermore, data on age and BMI (Body Mass Index), which belong to the basic data of the person being evaluated, are prepared. This basic data can be easily obtained from the results of a health checkup, or from hospital records. The FIM value estimation device inputs these simplified FIM values ​​and basic data into trained artificial intelligence, which then estimates FIM values ​​for 18 items and 7 values.

[0035] The FIM value estimation device outputs the estimated data of the FIM value calculated by the artificial intelligence on, for example, a display screen. In this way, the person evaluating the FIM value of the subject (evaluator) can easily obtain the FIM value of the subject for 18 items and 7 values. By trying various simplified FIM values, the inventors of the present application found that the simplified FIM value, which is expressed as three values ​​for the four items shown in Figure 1, can estimate the FIM value of 18 items and 7 values ​​with practical accuracy.

[0036] 2 is a block diagram illustrating the configuration of an FIM value estimation device that executes the data processing of FIG. 1. The FIM value estimation device 101 is a device that outputs estimated data of FIM values ​​for 18 items and 7 values ​​based on the simplified FIM value of a subject of FIM value assessment (subject to assessment) and basic data. In the illustrated example, the FIM value estimation device 101 is incorporated into a computer. That is, by installing and running a specific application on the computer, a processing device (processor) such as a central processing unit (CPU) of the computer functions as the FIM value estimation device 101.

[0037] The FIM value estimation device 101 in the illustrated example is connected to a network 5. In the illustrated example, the network 5 is the Internet. A server 7 owned by a facility such as a hospital and holding basic data such as the subject's medical records is connected to the network 5. The server 7 may be owned by an external vendor and used by the facility such as a hospital. The FIM value estimation device 101 can access the server 7 to obtain basic data such as the subject's age and BMI.

[0038] A server 9 that has constructed artificial intelligence that can be used through the network 5 may be connected to the network 5. The FIM value estimation device 101 uses artificial intelligence to estimate FIM values ​​for 18 items and 7 values ​​based on the simplified FIM value and basic data of the subject of FIM value assessment.

[0039] The artificial intelligence may be constructed in a computer as part of the FIM value estimation device 101 (artificial intelligence 23 in the illustrated example), or may be constructed in a computer separately from the FIM value estimation device 101 so as to be accessible by the FIM value estimation device 101, or may be artificial intelligence external to the computer, such as artificial intelligence provided by the server 9. In addition, a user terminal 3 may be connected to the network 5, which communicates with the FIM value estimation device 101 to operate the FIM value estimation device 101, input data to the FIM value estimation device 101, and obtain data output by the FIM value estimation device 101.

[0040] In the illustrated example, the FIM value estimation device 101 includes an interface 13, an input data receiving unit 15, a teacher data receiving unit 17, an estimation unit 19, a learning unit 21, an artificial intelligence 23, and an estimation data output unit 25. The interface 13 is a device part that enables communication between the FIM value estimation device 101 itself and external devices according to a predetermined protocol for each external device. Communication between the FIM value estimation device 101 and the user terminal 3, servers 7 and 9, an input device 27 such as a keyboard, an output device 29 such as a printer or display, and a storage medium 31 such as a USB memory or CD-ROM is performed through the interface 13. A program that causes a computer to function as the FIM value estimation device 101 may be supplied via the network 5 or by the storage medium 31 such as a CD-ROM.

[0041] 3 is a flowchart illustrating the processing flow of the FIM value estimation method realized by the FIM value estimation device 101. Note that the processing procedure of the FIM value estimation method illustrated in FIG. 3 can also be executed manually.

[0042] The input data accepting unit 15 accepts input data including the simplified FIM score and basic data of the person to be evaluated (step S1). The input data may be input from the input device 27, the storage medium 31, the server 7, or the user terminal 3. The estimation unit 19 inputs the input data accepted by the input data accepting unit 15 to the artificial intelligence 23, causing the artificial intelligence 23 to calculate estimated data for the FIM scores of 18 items and 7 values ​​(step S3).

[0043] If the artificial intelligence 23 has already been trained, it outputs highly accurate estimated data for the FIM values ​​of 18 items and 7 values. The estimated data output unit 25 outputs the estimated data calculated by the artificial intelligence 23 (step S5). The estimated data output by the estimated data output unit 25 is transmitted via the interface 13 to, for example, the user communication terminal 3 or the output device 29. This allows the person assessing the FIM values ​​to obtain the estimated data for the FIM values ​​of 18 items and 7 values.

[0044] If there is new input data and the process should not end (No in step S7), the process returns to step S1, and the input data accepting unit 15 accepts new input data (step S1). If the process should end (Yes in step S17), the process ends.

[0045] The computer incorporating the FIM value estimation device 101 may be, for example, one installed in an occupational therapist's facility, a care facility, a hospital, etc., or it may be a mobile computer that can be carried by an occupational therapist, caregiver, etc. to the home of the person being evaluated.

[0046] The artificial intelligence 23 is able to output highly accurate estimation data through machine learning. The FIM value estimation device 101 has a teacher data receiving unit 17 and a learning unit 21, which allows the FIM value estimation device 101 to train the artificial intelligence 23 by itself, without using an external artificial intelligence learning device. In other words, the FIM value estimation device 101 also has a built-in artificial intelligence learning device that trains the artificial intelligence 23 through machine learning.

[0047] Fig. 4 is a flowchart illustrating the flow of processing of the artificial intelligence learning method realized by the FIM value estimation device 101. Note that the processing procedure of the artificial intelligence learning method illustrated in Fig. 4 can also be executed manually.

[0048] When the FIM value estimation device 101 performs machine learning, the input data receiving unit 15 first receives input data including a simplified FIM value and basic data (step S61). The training data receiving unit 17 then receives training data, which are correct FIM values ​​of 18 items and 7 values ​​corresponding to the input data (step S63). Either process S61 or process S63 may be performed first, or they may be performed simultaneously.

[0049] Next, the learning unit 21 inputs the input data accepted by the input data accepting unit 15 and the teacher data accepted by the teacher data accepting unit 17 into the artificial intelligence 23, thereby training the artificial intelligence 23 to estimate the teacher data from the input data (step S65).

[0050] Next, when the FIM value estimation device 101 determines that the process should be repeated based on a user instruction or the like (Yes in step S67), it returns the process to S61. As a result, the input data receiving unit 15 receives new input data (step S61), and the teacher data receiving unit 17 receives new teacher data (step S63). When the FIM value estimation device 101 determines that the process should not be repeated (No in step S67), it ends the process. In this way, by inputting a large number of pairs of input data and teacher data that are associated with each other into the FIM value estimation device 101, the learning of the artificial intelligence 23 progresses and the accuracy of estimation improves.

[0051] In the past, simplified FIM values ​​and basic data collected from various subjects for evaluation, as well as FIM values ​​for 18 items and 7 values ​​obtained by actual measurements corresponding to these data, were associated with each other and recorded, for example, on storage medium 31. This allows the input data receiving unit 15 and the teacher data receiving unit 17 to sequentially read out a large number of data required for learning from storage medium 31, and the learning unit 21 to repeatedly learn the artificial intelligence 23 for each piece of data read out.

[0052] In this way, the FIM value estimation device 101 can switch between two operation modes: an estimation mode in which estimation data is calculated and output using the artificial intelligence 23, and a learning mode in which the artificial intelligence 23 undergoes machine learning. The switching of the operation modes can be instructed by, for example, the input device 27.

[0053] In the example of FIG. 2, the artificial intelligence 23 is incorporated into the computer 10 as part of the FIM value estimation device 101. Alternatively, as illustrated by the dotted line in FIG. 2, an artificial intelligence constructed on an external server 9 or the like may be used. In this case, the estimation unit 19, the learning unit 21, and the estimated data output unit 25 operate the external artificial intelligence via a network 5 or the like. The estimation unit 19 inputs input data accepted by the input data accepting unit 15 to the external trained artificial intelligence, causing it to output estimated data. The estimated data is sent to the estimated data output unit 25 via, for example, the interface 13. The estimated data output unit 25 outputs the received estimated data to an output device 29, a user communication terminal 3, or the like via the interface 13.

[0054] Furthermore, the learning unit 21 trains the external artificial intelligence to estimate the teacher data from the input data by inputting the input data received by the input data receiving unit 15 and the teacher data received by the teacher data receiving unit 17 to the external artificial intelligence. When the external artificial intelligence is used in this way, the artificial intelligence 23 that is a part of the FIM value estimation device 101 becomes unnecessary.

[0055] 5 is a schematic diagram illustrating the conceptual configuration of the artificial intelligence 23 used by the FIM value estimation device 101. The artificial intelligence provided by the server 9 also has a similar configuration, for example. The artificial intelligence 23 in the illustrated example is a neural network, and has an input layer 33 in which nodes that receive data are arranged, an output layer 37 in which nodes that output data resulting from calculations are arranged, and an intermediate layer 35 in which nodes connecting the input layer 33 and the output layer 37 are arranged. In the illustrated example, the intermediate layer 35 is a single layer, but it may be multiple layers. The value of the previous node is transmitted to the next node, reflecting the parameters assigned to each node, i.e., the weight and bias value of each node.

[0056] The input layer 33 receives input data received by the input data receiving unit 15, i.e., pairs of simplified FIM values ​​and basic data items. The input data is transmitted to the output layer 37 via the intermediate layer 35 while reflecting the parameters of each node. The data transmitted to the output layer 37 becomes estimated data for pairs of FIM values ​​(18 items) for 18 items and 7 values. The estimation unit 19 (see FIG. 2) inputs pairs of simplified FIM values ​​of the subject to be evaluated and basic data items to the input layer 33 of the artificial intelligence 23, and causes the output layer 37 to generate estimated data for pairs of FIM values ​​for 18 items and 7 values ​​of the subject to be evaluated. The estimated data output unit 25 outputs the generated estimated data after or without conversion, such as rounding.

[0057] In order for the estimated data appearing in the nodes of the output layer 37 to be highly accurate estimates of the FIM values ​​for 18 items and 7 values, it is necessary to train the artificial intelligence 23 using the actually measured FIM values ​​for 18 items and 7 values. Learning is performed by inputting a set of a simplified FIM value and basic data items for a given subject, the input of which is accepted by the input data accepting unit 15, to the input layer 33, and inputting training data for the same subject, i.e., a set of actually measured FIM values ​​for 18 items and 7 values, accepted by the training data accepting unit 17, to the output layer 37 as training data. The learning unit 21 (see FIG. 2) inputs this data to the artificial intelligence 23.

[0058] The artificial intelligence 23 calculates estimated data for the 18-item, 7-value FIM values ​​based on the input simplified FIM values ​​and basic data, generates the estimated data in the output layer 37, and calculates the error between the generated estimated data and the 18-item, 7-value FIM values ​​input as training data. The artificial intelligence 23 then changes the parameters of each node from the output layer 37 to the input layer 33, for example, using a well-known backpropagation algorithm, so that error-free estimated data is generated. This function is inherent to the artificial intelligence 23. By preparing many pairs of input data and training data and repeating learning, the artificial intelligence 23 can generate highly accurate estimated data. When training the artificial intelligence 23, it is also possible to adjust the number of intermediate layers 35 and the number of nodes in each layer to optimal values. Such techniques are well known.

[0059] Of the input data given to the artificial intelligence 23, the simplified FIM value is expressed as a three-level numerical value, for example, "1," "2," and "3." The basic data, age and BMI, are expressed as numerical values ​​based on those units, for example. When height and weight are included in the basic data, they can also be expressed as numerical values ​​based on those units. When gender is included in the basic data, they can be expressed as codes corresponding to male and female, for example, numerical values ​​such as "0" and "1." The output data, the FIM value of 18 items and 7 values, is expressed as a seven-level numerical value, for example, "1," "2," ... "7."

[0060] 5, other types of artificial intelligence may be used as the artificial intelligence 23, such as a decision tree-based LGBM (Light GBM; manufactured by Microsoft). LGBM has the advantage of being able to easily analyze which variables in the input data play an important role in the estimated data to be output, and is therefore particularly useful in the process of building the artificial intelligence 23.

[0061] The simplified FIM score belonging to the input data is generally not limited to the four items shown in Figure 1, but can also be fewer, more, or even less than 18 items. The more items there are, the higher the accuracy of the estimation, but the more preparation work is required. The four items shown in Figure 1 are two items belonging to the motor category: bed transfer and walking ability, and two items belonging to the cognitive category: problem-solving ability and memory. Even when selecting other items, it is desirable to include at least one of the 13 items belonging to the motor category and at least one of the five items belonging to the cognitive category as the simplified FIM score in terms of estimation accuracy.

[0062] The basic data included in the input data is not limited to the two items shown in Figure 1, and it is possible to select fewer, more, or different items. For example, in addition to age and BMI (body mass index), height and weight can also be added, and gender can also be added. Similarly, the more items there are, the higher the accuracy of the estimation will naturally be, but the amount of preparation work will also increase. It is also possible to remove the basic data from the input data and use only the simplified FIM value as input data. If the simplified FIM value items are shared and the basic data is removed from the input data, the accuracy of the estimation will decrease, but the amount of preparation work will be simpler.

[0063] Instead of or in addition to the basic data, it is also possible to add data on the subject's living environment, such as whether they are at home, hospitalized, or living in a nursing home, to the input data. Differences in living environment have a slight effect on the activities of daily living (ADL) represented by the 18-item, 7-value FIM score, even if the simplified FIM score or basic data is the same. Therefore, including living environment data in the input data contributes to improving the accuracy of estimation, just like including basic data.

[0064] Like the basic data, the living environment data can also be expressed numerically. For example, it can be expressed as a number such as "1," "2," or "3" depending on whether the person is at home, hospitalized, or living in a nursing home. Alternatively, it can be expressed as "1" if the person is at home, hospitalized, or living in a nursing home, and "0" if the person is not. In this case, the living environment data is expressed as a set of numbers such as "0,1,0." By converting the input data into numbers in this way, it becomes easier for the artificial intelligence 23 to handle the data.

[0065] The simplified FIM values ​​for four items and basic data for two items shown in Figure 1 were selected as relatively easy-to-prepare input data that play a major role in determining the FIM values ​​for 18 items and 7 values ​​as output data through the learning of artificial intelligence23 using LGBM, which is a decision tree, by preparing many pairs of input data and training data consisting of various items.In selecting the input data items, synthetic data, which is data generated by simulation, was referenced.

[0066] According to the test results, when the simplified FIM values ​​of the four items and the basic data of the two items shown in Figure 1 were used as input data, the accuracy of estimation was 73%, allowing for an error of ±1 above or below for the seven values. In other words, it was demonstrated that the input data shown in Figure 1 can be prepared relatively easily, and that the FIM values ​​of 18 items and 7 values ​​can be estimated with a practical level of accuracy. [Explanation of symbols]

[0067] 3 User communication terminal, 5 Network, 7, 9 Server, 13 Interface, 15 Input data reception unit, 17 Teacher data reception unit, 19 Estimation unit, 21 Learning unit, 23 Artificial intelligence, 25 Estimation data output unit, 27 Input device, 29 Output device, 31 Storage medium (memory), 33 Input layer, 35 Intermediate layer, 37 Output layer, 101 FIM value estimation device.

Claims

1. an input data receiving unit that receives input data including simplified FIM values ​​that represent the FIM values ​​of the subject of evaluation using a plurality of values ​​less than seven for fewer than 18 items; an estimation unit that inputs the input data received by the input data receiving unit into a trained artificial intelligence, and causes the artificial intelligence to calculate estimated data representing the FIM score of the subject of evaluation using seven values ​​for 18 items; and an estimated data output unit that outputs the estimated data calculated by the artificial intelligence.

2. The FIM value estimation device according to claim 1 , wherein the simplified FIM value includes at least one item out of 13 items belonging to motor items and at least one item out of five items belonging to cognitive items.

3. 3. The FIM value estimation device of claim 2, wherein the simplified FIM value includes two motor items, namely, bed transfer and walking, and two cognitive items, namely, problem-solving ability and memory, and each item is represented by three values.

4. The FIM value estimation device according to claim 1, wherein the input data includes basic data of the subject, including at least one item selected from the age, BMI (body mass index), height, and weight of the subject.

5. 2. The FIM value estimation device according to claim 1, wherein the input data includes living environment data of the person to be evaluated, and the living environment data is expressed by a numerical value.

6. an input data receiving process for receiving input data including simplified FIM values ​​in which the FIM values ​​of the subject are expressed by a plurality of values ​​less than seven for fewer than 18 items; an estimation process in which the input data received by the input data reception process is input to a trained artificial intelligence, and the artificial intelligence calculates estimated data representing the FIM score of the person to be evaluated using seven values ​​for 18 items; and an estimated data output process for outputting the estimated data calculated by the artificial intelligence.

7. The FIM value estimation method according to claim 6, wherein the simplified FIM value includes at least one item out of 13 items belonging to motor items and at least one item out of five items belonging to cognitive items.

8. The FIM value estimation method according to claim 7, wherein the simplified FIM value includes two motor items, namely, bed transfer and walking, and two cognitive items, namely, problem-solving ability and memory, and each item is represented by three values.

9. The FIM value estimation method according to claim 6, wherein the input data includes basic data of the subject, including at least one item selected from the age, BMI (body mass index), height, and weight of the subject.

10. 7. The FIM value estimation method according to claim 6, wherein the input data includes living environment data of the subject, and the living environment data is expressed by a numerical value.

11. 11. An FIM value estimation program that, when read by a computer, causes the computer to execute the FIM value estimation method according to any one of claims 6 to 10.

12. an input data receiving unit that receives input data including simplified FIM values ​​that represent the FIM values ​​of the subject of evaluation using a plurality of values ​​less than seven for fewer than 18 items; a teacher data receiving unit that receives teacher data corresponding to the input data and expressing the FIM values ​​of the subject of evaluation by seven values ​​for 18 items; An artificial intelligence learning device comprising: a learning unit that trains the artificial intelligence to estimate the teacher data from the input data by inputting the input data accepted by the input data accepting unit and the teacher data accepted by the teacher data accepting unit into the artificial intelligence.

13. 13. The artificial intelligence learning device according to claim 12, wherein the simplified FIM value includes at least one item out of 13 items belonging to motor items and at least one item out of 5 items belonging to cognitive items.

14. The artificial intelligence learning device of claim 13, wherein the simplified FIM value includes two items belonging to the motor category, namely, bed transfer and walking, and two items belonging to the cognitive category, namely, problem-solving ability and memory, and each item is represented by three values.

15. The artificial intelligence learning device according to claim 12, wherein the input data includes basic data of the person being evaluated, including at least one item selected from the age, BMI (body mass index), height, and weight of the person being evaluated.

16. 13. The artificial intelligence learning device according to claim 12, wherein the input data includes living environment data of the person to be evaluated, and the living environment data is expressed by a numerical value.

17. an input data receiving process for receiving input data including simplified FIM values ​​in which the FIM values ​​of the subject are expressed by a plurality of values ​​less than seven for fewer than 18 items; a teacher data receiving process for receiving teacher data corresponding to the input data, the teacher data representing the FIM values ​​of the subject of evaluation by seven values ​​for 18 items; An artificial intelligence learning method comprising: a learning process for inputting the input data accepted by the input data acceptance process and the teacher data accepted by the teacher data acceptance process into an artificial intelligence, thereby training the artificial intelligence to estimate the teacher data from the input data.

18. 18. The artificial intelligence learning method according to claim 17, wherein the simplified FIM value includes at least one item out of 13 items belonging to motor items and at least one item out of 5 items belonging to cognitive items.

19. The artificial intelligence learning method of claim 18, wherein the simplified FIM value includes two items belonging to the motor category, namely, bed transfer and walking, and two items belonging to the cognitive category, namely, problem-solving ability and memory, and each item is represented by three values.

20. The artificial intelligence learning method of claim 17, wherein the input data includes basic data of the person being evaluated, including at least one item selected from the age, BMI (body mass index), height, and weight of the person being evaluated.

21. The artificial intelligence learning method according to claim 17 , wherein the input data includes living environment data of the person to be evaluated, and the living environment data is expressed by a numerical value.

22. An artificial intelligence learning program that, when read by a computer, causes the computer to execute the artificial intelligence learning method according to any one of claims 17 to 21.

Images