Information processing system, information processing method, and control program for predicting adaptability to bruxism management method and indicating application range of said management method

Abstract

[Problem] To provide an information processing system with which it is possible to quantitatively and accurately predict adaptability to a bruxism management method and indicate an application range of a management method suitable for a target person. [Solution] The present invention comprises: a storage unit 31 for storing adaptability to a bruxism management method as an objective variable; a storage unit 32 for storing information pertaining to adaptability to pain, information related to intra-oral reflection, information related to cognitive behavior, information pertaining to concentration power, information pertaining to muscle tone, medical interview information, quality of sleep, past history, or numerical data pertaining to bruxism as explanatory variables; a processing unit 21 for distributing the objective variable and the explanatory variables to training data and verification data and storing said pieces of data; a processing unit 22 for constructing an adaptability prediction model using the objective variable and the explanatory variables through a machine learning method, and evaluating and reconstructing the prediction model using the objective variable and explanatory variables of the verification data; and a processing unit 23 for outputting the predicted adaptability.

Description

Information processing system, information processing method, and control program for predicting adaptability to bruxism management methods and indicating the scope of application of the management methods

[0001] The present invention relates to a method, system, and program for predicting adaptability to a bruxism management method using a machine learning model.

[0002] There are various management methods for treating, alleviating, etc. bruxism (involuntary contact between the upper and lower teeth, such as teeth grinding), including methods that detect bruxism and apply electrical or vibration stimulation (biofeedback methods, such as the bruxism management device "Butler GrindCare" provided by Sunstar Inc. (https: / / www.club-sunstar-pro.jp / product / detail / 261 and see Patent Document 1, etc.), methods that involve wearing a mouthpiece, physical therapy, and cognitive behavioral therapy. When these management methods are applied, a person's adaptability, that is, their tolerance for pain, etc. and the degree of effectiveness of treatment and alleviation, varies greatly from person to person, and there is particularly large individual variation in adaptability when biofeedback (stimulation of the human body) methods are applied.

[0003] Therefore, in the past, it was difficult to predict in advance the suitability of a person for a bruxism management method (the above-mentioned tolerance and degree of effectiveness), and it was not always possible to provide a bruxism management method that was suitable for that person (subject).

[0004] JP 2015-33520 A (Japan)

[0005] Therefore, in consideration of the above-mentioned situation, the present invention aims to solve the problem of providing an information processing system that can quantitatively and accurately predict the adaptability of bruxism management methods (tolerance and effects of treatment / alleviation), and can indicate to the target of treatment / alleviation and dentists, etc., the range of application of management methods that are suitable for the target.

[0006] That is, the present invention includes the following inventions. (1) An information processing system for predicting adaptability to a bruxism management method and indicating the scope of application of the management method, the information processing system comprising: an objective variable memory unit that stores information on adaptability to the bruxism management method as an objective variable; an explanatory variable memory unit that receives input of at least one of information related to the person applying the bruxism management method, information related to oral reflexes, information related to cognitive behavior, information related to concentration, information related to muscle tension, information on bruxism interviews, information on sleep quality, medical history related to bruxism management, and numerical data on bruxism, and stores the information on the objective variable and explanatory variables as explanatory variables; an allocation processing unit that sorts and stores the information on the objective variable and explanatory variables into training data and verification data; a prediction model construction processing unit that uses a machine learning technique to construct a prediction model that predicts adaptability to the bruxism management method using the objective variable and explanatory variables of the training data, evaluates the prediction model using the objective variable and explanatory variables of the verification data, and reconstructs the prediction model as necessary; and an output processing unit that outputs the adaptability to the bruxism management method using the prediction model.

[0007] (2) The information processing system according to (1), wherein the predictive model construction processing unit evaluates the predictive accuracy of the predictive model and the presence or absence of overfitting using the objective variable and explanatory variables for the validation.

[0008] (3) The information processing system according to (2), wherein the prediction model construction processing unit evaluates the prediction accuracy and determines whether overlearning has occurred based on an R value and an RMSE value calculated by the following formulas:

[0009]

[0010] There are no clear rules for the R2 value, but it is considered good to have an R2 value of 0.8 or higher. RMSE is the square root of the root mean square of the actual value minus the predicted value. It is used to compare training and validation data, and the smaller the value, the better. For example, if the R2 value is "0.889" for the training data and "0.372" for the validation data, the model is evaluated as having overfitted and in need of improvement.

[0011] (4) The information processing system according to (3), wherein the sorting unit further sorts the validation data into two types: internal validation data and external validation data; and the predictive model construction unit uses the internal validation data to determine whether overfitting has occurred and uses the external validation data to evaluate prediction accuracy.

[0012] (5) The information processing system according to (1), wherein the allocation processing unit compresses each piece of information about the objective variable and the explanatory variable by normalizing and standardizing the information to match the scale (number of digits), and then allocates the information to the training data and the validation data and stores the data.

[0013] (6) A control program for causing a computer to function as the information processing system described in (1), the control program causing a computer to function as the allocation processing unit, the prediction model construction processing unit, and the output processing unit.

[0014] (7) A storage medium that stores a control program for causing a computer to function as the information processing system described in (1), the storage medium storing a control program for causing a computer to function as the allocation processing unit, the prediction model construction processing unit, and the output processing unit.

[0015] (8) An information processing method executed by a system that predicts adaptability to a bruxism management method and indicates the scope of application of the management method, the system having an objective variable memory unit that stores information on adaptability to a bruxism management method as an objective variable, and an explanatory variable memory unit that receives input of at least one of information related to the adaptability to pain of a person applying the bruxism management method, information related to intraoral reflexes, information related to cognitive behavior, information related to concentration, information on muscle tension, information on a medical interview regarding bruxism, information on sleep quality, medical history related to bruxism management, and numerical data on bruxism, and stores the information as explanatory variables, the information processing method including: a sorting step that sorts and stores each piece of information on the objective variable and the explanatory variable into training data and verification data; a predictive model construction step that uses a machine learning technique to construct a predictive model that predicts adaptability to the bruxism management method using the objective variables and explanatory variables of the training data, evaluates the predictive model using the objective variables and explanatory variables of the verification data, and reconstructs the predictive model as necessary; and an output step that outputs the adaptability to the bruxism management method using the predictive model.

[0016] (9) The information processing method according to (8), wherein the prediction model construction step evaluates the prediction accuracy of the prediction model and the presence or absence of overfitting using the objective variable and explanatory variables for the validation.

[0017] (10) The information processing method according to (9), wherein the prediction model construction step evaluates the prediction accuracy and determines whether overlearning has occurred based on an R value and an RMSE value calculated by the following formulas:

[0018] (11) The information processing method according to (10), wherein the sorting step further sorts the validation data into two types: internal validation data and external validation data; and the prediction model building step determines whether or not overfitting has occurred using the internal validation data and evaluates prediction accuracy using the external validation data.

[0019] (12) The information processing method according to (8), wherein the allocation step includes compressing the information of the objective variable and the explanatory variable by normalizing and standardizing the information to match the scale (number of digits), and then allocating the information to the training data and the validation data and storing the data.

[0020] According to the present invention as described above, it is possible to provide an information processing system that can quantitatively and accurately predict the adaptability of bruxism management methods (tolerance and effects of treatment / alleviation), and can indicate to the target of treatment / alleviation and dentists, etc., the range of application of management methods that are suitable for the target.

[0021] FIG. 1 is a block diagram showing a schematic configuration of an information processing system according to a representative embodiment of the present invention. FIG. 2 is an explanatory diagram explaining the contents of a target variable storage unit of the information processing system. FIG. 3 is an explanatory diagram explaining the contents of an explanatory variable storage unit of the information processing system. FIG. 4 is an explanatory diagram showing an example of a check box for an evaluation method (for listening) of adaptability to pain. (a) and (b) are explanatory diagrams each showing an example of an output display. (b) is an explanatory diagram showing an example of an output display.

[0022] Next, an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

[0023] The information processing system according to the present invention is an information processing system that quantitatively predicts the adaptability of a bruxism management method using a computer, and is composed of one or more information processing devices, each including a processing device 2, a storage device 3, an information display unit 4, and an input device 5, as shown in FIG. 1. Specifically, the information processing system is a computer device that includes the processing device 2 as the core, the storage device 3, the input device 5 such as a pointing device, keyboard, or touch panel, the information display unit 4 such as a display, and other components such as a communication control unit (not shown). The input device 5 and the information display unit 4 are not essential. The information processing system may be communicatively connected to a terminal computer owned by a subject or dentist, and information may be input from the terminal computer or transmitted to the terminal computer and output on the display unit of the terminal computer. The information processing system may also be communicatively connected to an information server of an external institution, and information from the server's database may be input.

[0024] The processing device 2 is mainly composed of a CPU such as a microprocessor, and has a storage unit consisting of RAM and ROM (not shown) in which programs defining the procedures of various processing operations and processing data are stored. The storage means 3 consists of a memory or hard disk inside or outside the information processing system 1. Some or all of the contents of the storage unit may be stored in the memory or hard disk of another computer connected to the information processing system 1 for communication.

[0025] The storage means 3 has a dependent variable storage section 31 that stores information on adaptability to the bruxism management method as a dependent variable, and an explanatory variable storage section 32 that stores at least one of information related to the adaptability to pain of the person applying the bruxism management method, information related to intraoral reflex (gag reflex), information related to cognitive behavior, information related to concentration, information on muscle tension, information on medical interview regarding bruxism, information on sleep quality, medical history related to bruxism management, and numerical data on bruxism as explanatory variables.

[0026] The objective variable (information on adaptability to bruxism management methods) can be information from medical institutions, websites and papers of bruxism management research institutions, information from the information databases of each institution, etc., which can be input using the input means 5, etc. Specific contents include adaptability data 31a such as the degree of effectiveness, tolerance to pain, etc., and suitability of various biofeedback management methods (stimulation of a person by electricity, vibration, sound, etc.), adaptability data 31b such as the degree of effectiveness, information on oral reflexes and vomiting, etc., and suitability of a mouthpiece management method, adaptability data 31c such as the degree of effectiveness, tolerance to pain, etc., and suitability of physical therapy, and adaptability data 31d such as the degree of effectiveness, information on cognitive behavior and concentration, etc., and suitability of a cognitive behavioral therapy management method, etc.

[0027] The explanatory variables are input via the input means 5 or the like, and, as shown in Fig. 3, preferred examples include pain tolerance 32a (inputting different electrical or vibration stimulation levels and inputting the tolerance and judgment results for each stimulation level) 32b, oral reflex tolerance (inputting the judgment results for the oral reflex level using a medical interview or an instrument or a sensor device) 32b, and cognitive behavior tolerance (inputting the judgment results for the concentration level using a medical interview or an instrument or a sensor device) 32c. It is preferable to use at least one of these as an essential explanatory variable.

[0028] The methods for evaluating and measuring adaptability to pain will now be described in more detail. Pain can be measured using, for example, the visual analogue scale (VAS), numerical rating scale (NSR), verbal rating scale (VRS), face rating scale (FRS), etc., as sensitivity to pain when an electrical stimulus is actually applied. In addition, a PainVision quantitative pain sensory analyzer can be used as a medical device for calculating the degree of pain numerically.

[0029] Regarding the type of pain felt, it is possible to ask the patient, for example, whether it was a pin-and-needle type pain, an electric shock type pain, a burning or tingling pain, a numbing pain, or no pain at all, as shown in Figure 4. Furthermore, the pain level is preferably calculated by {(pain-related current value) - (minimum perceptible current value)} / (minimum perceptible current value) x 100. Measurement sites are preferably the temples, chin, forearms, etc., and the vibration and electrical stimulation preferably use frequencies and current values ​​actually used in bruxism management methods.

[0030] Also preferred are information 32d on the results of a medical interview regarding bruxism, information 32e on the medical history of bruxism management using the mouthpiece method / physical therapy (massage, etc.) (including information on suitability such as the degree of effectiveness, tolerance to pain, etc., and suitability), the number of bruxism episodes per hour (information obtained using a device with a sensor) 32f, information on sleep (including information on the quality and depth of sleep (information obtained using a device with a sensor (information from a smartphone app, etc.) and information on the results of the medical interview) 32g, and information on muscle tension, etc. (including information obtained using a device with a sensor and information on the results of the medical interview and palpation) 32h. It is preferable to selectively use one or more of these as explanatory variables.

[0031] The information on the intraoral reflex level can be determined by, for example, interviewing the patient and classifying the degree of nausea (vomiting) caused by toothpaste entering the mouth during tooth brushing into the following levels G1 to G5. Note that this can be the result of an examination by a dentist or the result of the patient's own assessment.

[0032] (Levels) ・"G5: Severe intraoral reflex" When a gag reflex occurs when a dental mirror is temporarily inserted into the mouth (without touching the oral mucosa) ・"G4: Strong intraoral reflex" When a gag reflex occurs when the alveolar mucosa of the anterior teeth is touched with a dental mirror ・"G3: Moderate intraoral reflex" When a gag reflex occurs when the alveolar mucosa of the molars is touched ・"G2: Weak intraoral reflex" When a gag reflex occurs when the center of the posterior edge of the posterior palate is touched ・"G1: Not weak, but within normal range" When a gag reflex occurs when the base of the tongue is touched ・"G0: ​​No intraoral reflex" When a gag reflex does not occur no matter where the mouth is touched

[0033] The information on cognitive behavior and concentration includes, for example, information obtained by interviews regarding the degree of self-management, self-monitoring ability, willingness to learn, and level of motivation (willingness to improve on events), information obtained by psychological tests such as the Stroop test that measures attention and cognitive control, and information obtained from self-management data such as a concentration diary that records daily concentration, and it is preferable to determine each level. Note that it may be the results of an examination by a dentist or the results of self-determination.

[0034] Functionally, the processing device 2 comprises an allocation processing unit 21 that allocates and stores each piece of information on the objective variables and explanatory variables into learning data and verification data, a prediction model construction processing unit 22 that uses a machine learning technique to construct a prediction model of adaptability to the bruxism management method using the objective variables and explanatory variables of the learning data, evaluates the prediction model using the objective variables and explanatory variables of the verification data, and performs any necessary reconstruction of the prediction model, and an output processing unit 23 that uses the prediction model to output the predicted adaptability to the bruxism management method (tolerance and effects of treatment and alleviation) to the information display unit 4, etc., and these processing functions are realized by the above-mentioned program.

[0035] The validation data allocated by the allocation processing unit 21 is test data for evaluating whether the constructed prediction model can be applied to unknown data. When allocating the training data and validation data, it is preferable to perform a process to align the scale by compressing it to 0 to 1 through normalization and standardization. This allows data with different digits to be handled using the same calculation formula. The allocation processing unit 21 also further allocates the validation data into two types: internal validation data and external validation data. Specifically, the information on the objective variable and explanatory variables is compressed through normalization and standardization to align the scale (number of digits), and then allocated and stored as training data and validation data.

[0036] The predictive model construction processing unit 22 constructs a model based on the training data, avoids overfitting by adopting a model that fits well to the internal validation data, and objectively evaluates the prediction accuracy using the external validation data. Predictive models can be constructed using ANN, Read-across, and the like. Among these, the CatBoost model is a GBDT algorithm published in 2017 that effectively handles categorical variables using a permutation-driven method and is less prone to overfitting. Because the present invention handles several categorical variables as explanatory variables, it is preferable to use CatBoost.

[0037] The prediction model construction processing unit 22 evaluates the prediction accuracy of the prediction model and the presence or absence of overfitting using the objective variable and explanatory variables for verification. Specifically, the prediction accuracy is evaluated and the presence or absence of overfitting is determined using the R value and RMSE value calculated by the following formulas.

[0038]

[0039] There are no clear regulations regarding the R2 value, but a value of 0.8 or higher is considered good. RMSE is the square root of the root mean square of the actual value minus the predicted value. It is used to compare training and validation data, and the smaller the value, the better. For example, if the R2 value is "0.889" for the training data and "0.372" for the validation data, the model is evaluated as having overfitted and in need of improvement. The presence or absence of overfitting is determined using internal validation data. Prediction accuracy is evaluated using external validation data.

[0040] The output processing unit 23 displays and outputs the results of predicted adaptability (tolerance and therapeutic / alleviation effects) to the bruxism management method on the information display unit 4 such as a display, and transmits the results to a client computer connected for communication. It is preferable to display the adaptability (tolerance and therapeutic / alleviation effects) to the subject for each management method, such as the biofeedback method, the mouthpiece method, or physical therapy, as a numerical index.

[0041] Specifically, as shown in Figure 5(a), for example, the adaptability evaluation can be displayed as a percentage out of a total of 100% for four types of methods: mouthpiece method, biofeedback method, physical therapy, and cognitive behavioral therapy. The adaptability may also be displayed on a 10-point scale instead of a percentage. In the example of Figure 5(a), the mouthpiece method is rated the highest at 50%, and it can be immediately seen that it has the best adaptability.

[0042] It is also preferable to display the compatibility in easy-to-understand words using five levels: "compatible," "fairly compatible," "average," "not very compatible," and "uncompatible." The number of levels is not limited to five, and three or ten levels may be used. In the example of FIG. 5(b), it can be immediately seen that the biofeedback method and physical therapy have excellent adaptability. In addition to numbers and words, it is also preferable to display the compatibility using smiley faces as shown in FIG. 6, or using color correlation (heat map).

[0043] It is also preferable to list the bruxism management methods for the subject in order of their applicability, and in this embodiment, this is indicated as a prompt, and an output processing unit 23 having a generation AI (large-scale language model) outputs a list of the applicability evaluation results to the information display unit 4, etc.

[0044] Furthermore, in order to improve the performance of the adaptability evaluation results through learning feedback, it is preferable to obtain actual user usage results as input. Examples of manual input (subjective data on the difference from expectations and whether there is improvement) include sleep quality, discomfort around the oral cavity, stress and anxiety levels, and symptoms such as pain. It is also preferable to input data comparing the number of times teeth grinding occurred before and after using a sensor such as a surface electromyogram (EMG) sensor, or to input sleep quality data and stress and anxiety levels measured by a smartwatch or electrocardiogram (ECG) sensor.

[0045] Although the embodiments of the present invention have been described above, the present invention is not limited to these embodiments. For example, instead of configuring the processing device using software processing by a computer, it is preferable to configure part or all of it using hardware processing circuits. In this case, an artificial intelligence processing circuit can also be used as the machine learning mechanism. It goes without saying that the present invention can be embodied in various forms within the scope of the gist of the present invention.

[0046] REFERENCE SIGNS LIST 1 Information processing system 2 Processing device 3 Storage means 4 Information display unit 5 Input means 21 Allocation processing unit 22 Prediction model construction processing unit 23 Output processing unit 31 Objective variable storage unit 32 Explanatory variable storage unit

Claims

1. An information processing system that predicts adaptability to a bruxism management method and indicates the scope of application of the management method, comprising: an objective variable memory unit that stores information on adaptability to a bruxism management method as an objective variable; an explanatory variable memory unit that receives input of at least one of the following information about the person applying the bruxism management method: information related to their adaptability to pain, information related to intraoral reflexes, information related to cognitive behavior, information related to concentration, information on muscle tension, information about bruxism interviews, information about sleep quality, medical history related to bruxism management, and numerical data about bruxism, and stores the information as explanatory variables; a sorting processing unit that sorts the information about the objective variable and explanatory variables into training data and verification data and stores the data; a prediction model construction processing unit that uses a machine learning technique to construct a prediction model that predicts adaptability to the bruxism management method using the objective variable and explanatory variables of the training data, evaluates the prediction model using the objective variable and explanatory variables of the verification data, and reconstructs the prediction model as necessary; and an output processing unit that outputs the adaptability to the bruxism management method using the prediction model.

2. The information processing system according to claim 1, wherein the predictive model construction processing unit uses the objective variable and explanatory variables for the validation to evaluate the predictive accuracy of the predictive model and whether or not overfitting has occurred.

3. The information processing system according to claim 2, wherein the prediction model construction processing unit evaluates the prediction accuracy and determines whether overlearning has occurred based on an R2 value and an RMSE value calculated by the following formulas:

4. The information processing system of claim 3, wherein the sorting processing unit further sorts the verification data into two types: internal verification data and external verification data; and the predictive model construction processing unit uses the internal verification data to determine whether overfitting has occurred and uses the external verification data to evaluate prediction accuracy.

5. The information processing system of claim 1, wherein the allocation processing unit compresses the information of the objective variables and explanatory variables by normalizing and standardizing them to match the scale (number of digits), and then allocates and stores the data as training data and verification data.

6. A control program for causing a computer to function as the information processing system of claim 1, the control program causing a computer to function as the allocation processing unit, the prediction model construction processing unit, and the output processing unit.

7. A storage medium for storing a control program for causing a computer to function as the information processing system described in claim 1, the storage medium storing a control program for causing a computer to function as the allocation processing unit, the prediction model construction processing unit, and the output processing unit.

8. An information processing method executed by a system that predicts adaptability to a bruxism management method and indicates the scope of application of the management method, the system having an objective variable memory unit that stores information on adaptability to a bruxism management method as an objective variable, and an explanatory variable memory unit that receives input of at least one of information related to the adaptability to pain of a person applying a bruxism management method, information related to intraoral reflexes, information related to cognitive behavior, information related to concentration, information on muscle tension, information on a medical interview regarding bruxism, information on sleep quality, medical history related to bruxism management, and numerical data on bruxism, and stores the information as explanatory variables, the information processing method comprising: a sorting step of sorting and storing each piece of information on the objective variable and explanatory variables into training data and verification data; a predictive model construction step of constructing a predictive model that predicts adaptability to the bruxism management method using the objective variables and explanatory variables of the training data by a machine learning technique, evaluating the predictive model using the objective variables and explanatory variables of the verification data, and reconstructing the predictive model as necessary; and an output step of outputting the adaptability to the bruxism management method using the predictive model.

9. The information processing method according to claim 8, wherein the predictive model construction step uses the objective variable and explanatory variables for validation to evaluate the predictive accuracy of the predictive model and whether or not overfitting has occurred.

10. The information processing method according to claim 9, wherein the prediction model construction step evaluates the prediction accuracy and determines whether overlearning has occurred based on an R2 value and an RMSE value calculated by the following formulas.

11. An information processing method according to claim 10, wherein the sorting procedure further sorts the validation data into two types: internal validation data and external validation data, and the predictive model construction procedure uses the internal validation data to determine whether overfitting has occurred and uses the external validation data to evaluate prediction accuracy.

12. An information processing method according to claim 8, wherein the allocation procedure compresses the information of the objective variables and explanatory variables by normalizing and standardizing them to match the scale (number of digits), and then allocates and stores the data as training data and validation data.

Images