KR102478102B1 - Method of treatment for insomnia - Google Patents

Abstract

A method of treating insomnia provided by a computer is presented. The method for treating insomnia of the present invention inputs sleep-related information to a pre-learned algorithm as an input value, and obtains a first sleep efficiency index representing objective sleep efficiency and a second sleep efficiency index representing subjective sleep efficiency as output values and providing a treatment method related to sleep based on the first sleep efficiency index and the second sleep efficiency index.

Description

Insomnia treatment method {METHOD OF TREATMENT FOR INSOMNIA}

The present invention relates to a method for treating insomnia.

Insomnia generally refers to a sleep disorder, which includes insomnia that makes it difficult to fall asleep, sleep maintenance disorder in which you often wake up during sleep or wake up too early even though you fall asleep, and sleep that is sufficient for you to feel even though you have slept relatively much. Cognitive impairment, in which a person feels that they have not achieved

Sleep occupies about 1/3 of human life and plays an important role in energy storage and recovery as well as hormone secretion and memory consolidation. If you do not get enough sleep or do not get a good night's sleep, your mental activity becomes blurred and your body dulls. In modern society, an increasing number of people complain of sleep disorders due to environmental factors such as aging of the population, increased stress, aging, and changes in sleep cycles.

Furthermore, sleep has been reported to be associated with psychiatric disorders such as depression. For example, if you have symptoms of a mental illness or if it becomes severe, your sleep quality may deteriorate. In addition, when the degree of mental illness is improved, an improvement in sleep state may appear as an indicator for this. As such, the sleep state can be used as important information related to the degree, prognosis, or treatment effect of a mental disorder.

Therefore, in the treatment of sleep disorders, it may be important to accurately analyze the sleep state of an individual and to identify factors that affect sleep.

On the other hand, since sleep disorders occur while sleeping, it is difficult for even the person to identify the symptoms. In addition, polysomnography, which comprehensively examines brain waves, muscle movements, breathing, electrocardiogram, oxygen saturation, etc. during sleep, can be used to accurately diagnose sleep disorders. is a difficult situation. Furthermore, such a sleep analysis method is not suitable for home use due to uncertainties due to errors, and a number of electrodes must be attached to the user's head.

Accordingly, the equipment-based sleep analysis method may impose a great burden on the patient. Furthermore, these factors can be recognized as an unstable experience by a significant number of patients in sleep analysis, which can affect the analysis result.

Accordingly, there is a demand for the development of a system capable of overcoming the above limitations and effectively analyzing the quality of sleep. In particular, the development of a sleep analysis method and device for analyzing the quality of sleep for a user with a sleep disorder, analyzing the cause thereof, and providing it to the user can be very important in preventing, diagnosing, and even treating sleep disorders. there is.

Republic of Korea Patent Registration No. 10-2053604 (2019.12.09. notice)

One object of the present invention is to measure sleep efficiency in an optimized way, and to provide an optimized method for treating insomnia based thereon.

Another object of the present invention is to provide a method for treating insomnia that can improve insomnia by improving objective sleep efficiency and subjective sleep efficiency and reducing the difference between objective sleep efficiency and subjective sleep efficiency.

Another object of the present invention is to provide an insomnia treatment method capable of simultaneously detecting objective/subjective sleep efficiency while preventing overfitting in a time series using a plurality of sleep indicators.

The problems to be solved by the present invention are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below.

A method for treating insomnia provided by a computer according to an embodiment of the present invention for solving the above problems is a first sleep efficiency index representing objective sleep efficiency by inputting information related to sleep to a previously learned algorithm as an input value. and obtaining a second sleep efficiency index representing subjective sleep efficiency as an output value and providing a treatment method related to sleep based on the first sleep efficiency index and the second sleep efficiency index.

In an embodiment, the information related to sleep may include information measured through at least one of a mobile terminal and a wearable device and input information for a pre-written question.

In an embodiment, the pre-learned algorithm is characterized in that an LSTM algorithm for performing time series modeling and a random forest algorithm capable of preventing overfitting are fused.

In an embodiment, the pre-learned algorithm is characterized in that in a random forest algorithm using a decision tree, an LSTM algorithm is used instead of the decision tree.

In an embodiment, the pre-learned algorithm is characterized in that it performs multi-tasking configured to output a plurality of output values.

In an embodiment, the obtaining step is characterized in that the first sleep efficiency index and the second sleep efficiency index are output through multi-tasking applied to the pre-learned algorithm, and the first sleep efficiency index and The method further includes calculating a third sleep efficiency index corresponding to a difference between the second sleep efficiency index.

In an embodiment, a predetermined number of first sleep index related to the first sleep efficiency index and second sleep index related to the second sleep efficiency index are extracted from among a plurality of sleep indexes included in the sleep related information. Including more steps to

In an embodiment, it is characterized in that the extracted first sleep index and the second sleep index are upper predetermined number of sleep indexes having a great influence on the first sleep efficiency index and the second sleep efficiency index, respectively. .

In an embodiment, the providing may include providing a treatment method related to at least one of the first sleep index and the second sleep index.

In an embodiment, the step of collecting new sleep-related information after applying the provided treatment method, and inputting the new sleep-related information and the provided treatment method as input values of the previously learned algorithm, The method further includes acquiring a sleep efficiency index and a new second sleep efficiency index as output values.

A computer program providing a method for treating insomnia according to another embodiment of the present invention for solving the above problems is combined with a computer that is hardware and stored in a medium to perform any one of the above methods.

In addition to this, another method for implementing the present invention, another system, and a computer readable recording medium recording a computer program for executing the method may be further provided.

According to the present invention, sleep efficiency can be more accurately measured without using expensive equipment.

In addition, according to the present invention, it is possible to simultaneously calculate objective sleep efficiency and subjective sleep efficiency by inputting high-dimensional sleep-related variables into a time-series and non-overfitting model, thereby obtaining accuracy and speed for measuring sleep efficiency at the same time. can

In addition, according to the present invention, it is possible to provide a more direct and effective treatment method for insomnia by extracting sleep indicators that greatly affect objective sleep efficiency and subjective sleep efficiency and providing a treatment method related thereto.

The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

1 is a flow chart for explaining a representative insomnia treatment method of the present invention.
2 is a conceptual diagram for explaining objective sleep efficiency and subjective sleep efficiency of the present invention.
3 is a conceptual diagram for explaining a pre-learned algorithm for calculating a first sleep efficiency index representing objective sleep efficiency and a second sleep efficiency index representing subjective sleep efficiency.
4 is a conceptual diagram for explaining multi-tasking applied to the pre-learned algorithm of the present invention.
5 is a conceptual diagram for explaining sleep indexes affecting first and second sleep efficiency indices in the present invention.

Advantages and features of the present invention, and methods of achieving them, will become clear with reference to the detailed description of the following embodiments taken in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, only these embodiments are intended to complete the disclosure of the present invention, and are common in the art to which the present invention belongs. It is provided to fully inform the person skilled in the art of the scope of the invention, and the invention is only defined by the scope of the claims.

Terminology used herein is for describing the embodiments and is not intended to limit the present invention. In this specification, singular forms also include plural forms unless specifically stated otherwise in a phrase. As used herein, "comprises" and/or "comprising" does not exclude the presence or addition of one or more other elements other than the recited elements. Like reference numerals throughout the specification refer to like elements, and “and/or” includes each and every combination of one or more of the recited elements. Although "first", "second", etc. are used to describe various components, these components are not limited by these terms, of course. These terms are only used to distinguish one component from another. Accordingly, it goes without saying that the first element mentioned below may also be the second element within the technical spirit of the present invention.

Unless otherwise defined, all terms (including technical and scientific terms) used in this specification may be used with meanings commonly understood by those skilled in the art to which the present invention belongs. In addition, terms defined in commonly used dictionaries are not interpreted ideally or excessively unless explicitly specifically defined.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Prior to the description, the meaning of the terms used in this specification will be briefly described. However, it should be noted that the description of terms is intended to help the understanding of the present specification, and is not used in the sense of limiting the technical spirit of the present invention unless explicitly described as limiting the present invention.

In this specification, 'computer' includes all various devices that can perform calculation processing and provide results to users. For example, a computer includes not only a desktop PC and a notebook (Note Book) but also a smart phone, a tablet PC, a cellular phone, a PCS phone (Personal Communication Service phone), synchronous/asynchronous A mobile terminal of IMT-2000 (International Mobile Telecommunication-2000), a Palm Personal Computer (Palm PC), and a Personal Digital Assistant (PDA) may also be applicable. In addition, the computer may correspond to a server that receives a request from a client and performs information processing.

The insomnia treatment method described in this specification may be performed by the computer described above.

1 is a flow chart for explaining a representative insomnia treatment method of the present invention.

First, in the insomnia treatment method provided by a computer according to an embodiment of the present invention, the computer inputs information related to sleep to a pre-learned algorithm as an input value, and a first sleep efficiency index representing objective sleep efficiency and subjective sleep efficiency A step of obtaining the second sleep efficiency index indicating efficiency as an output value proceeds (S110).

Thereafter, the computer proceeds to provide a treatment method related to sleep based on the first sleep efficiency index and the second sleep efficiency index (S120).

2 is a conceptual diagram for explaining objective sleep efficiency and subjective sleep efficiency of the present invention.

First, the computer of the present invention may measure and receive sleep-related information (or sleep index), and calculate Sleep Efficiency (SE) based thereon.

Here, sleep efficiency may include objective sleep efficiency (Objective SE) and subjective sleep efficiency (Subjective SE).

The computer may calculate a first sleep efficiency index (Y1) representing objective sleep efficiency and a second sleep efficiency index (Y2) representing subjective sleep efficiency, based on the information related to sleep.

Here, the information related to sleep may mean a plurality of sleep indexes used to calculate the first and second sleep efficiency indices.

In this case, as shown in FIG. 2 , the sleep-related information may include information measured through at least one of a mobile terminal and a wearable device and input information for a pre-written question.

Referring to FIG. 2, information measured through at least one of a mobile terminal and a wearable device includes sleep start/end time, total sleep time, REM/light/deep sleep, awakening time during sleep, snoring data, oxygen saturation, It may include sleep indicators objectively measured through sensors provided in mobile terminals and wearable devices, such as heart rate, respiration rate, amount of activity, and intensity of illumination.

In addition, input information for a pre-written question means information input by a user with respect to a pre-written question, which may mean a subjective sleep index. For example, the input information for the prepared questions may include answers to various questions related to sleep, such as a stress scale, a depression/anxiety scale, a clinical outcome scale, or a clinical questionnaire.

The first sleep efficiency index representing objective sleep efficiency may be a value expressed as a percentage of the time actually slept compared to the time actually lying in bed.

In addition, the second sleep efficiency index, which represents subjective sleep efficiency, is not the time the user actually slept compared to the time the user was lying down, but how much sleep the user feels subjectively felt compared to the time the user was lying down. It can mean the indicated value.

Insomnia described in this specification is not only when the first sleep efficiency index and the second sleep efficiency index are low, but also when the difference (Y3) between the first sleep efficiency index and the second sleep efficiency index is large (ie, actually relatively You may have gotten enough sleep, but you feel that you haven't gotten enough sleep).

The insomnia treatment method of the present invention is not only a treatment method to increase the first sleep efficiency index (Y1) and the second sleep efficiency index (Y2), but also the first sleep efficiency index (Y1) and the second sleep efficiency index (Y2) ) can also provide a treatment method that reduces the difference (Y3).

On the other hand, the insomnia treatment method of the present invention does not simply calculate the first and second sleep efficiency indices directly through a wearable device or information input by a user, but through an algorithm to which machine learning is applied. First and second sleep efficiency indices may be calculated.

Specifically, as described above, the computer may input information related to sleep as an input value of the pre-learned algorithm and obtain first and second sleep efficiency indices as output values.

Such sleep-related information includes various information measured by wearable devices (objective sleep indicators including sleep start/end time, oxygen saturation, heart rate, etc.), subjective sleep-related indicators including stress scales, depression/anxiety scales, etc. It can include bio-signal data, polysomnography (PSG) including EEG, EMG, etc., and individual basic information including gender, age, and disease status, and these information will be named sensor/life-log-based data. can

The pre-learned algorithm may be learned in various ways.

For example, the device (computer) providing the insomnia treatment method of the present invention can collect data from a specific user (specific insomnia patient) for a certain period of time, and learn an algorithm for each individual based on this.

This may be understood as a concept of learning a model to which machine learning is applied for each patient in order to calculate a treatment method optimized for each patient.

In order to learn a model (algorithm) applicable to each patient, the device that provides the insomnia treatment method provides the treatment method in a general guide (Rule-based) until the data for each patient is accumulated for more than a predetermined number of days, When the data of is accumulated for more than the predetermined number of days, an algorithm can be learned through the accumulated data to generate a learned algorithm capable of providing a treatment method.

As another example, a device (computer) providing a method for treating insomnia may learn a generally applicable algorithm (model) based on data obtained from a plurality of people.

That is, the computer learns an algorithm (model) capable of providing an insomnia treatment method using data obtained through a plurality of people, and then, through data applied to each person, an algorithm optimized for each person ( The algorithm (model) may be tuned so that the model) is applied.

In order to tune a general model (general algorithm) learned based on data acquired through a plurality of people to be optimized for each individual, the computer uses a model (algorithm) learned based on data obtained from a plurality of people (all patients). It is possible to proceed with a tuning process that is specific to a particular patient (individually).

In order to proceed with this tuning process, the computer additionally learns the data acquired for a predetermined number of days for each individual to the algorithm learned (trained) based on the data of a plurality of people, and the algorithm tuned for each individual (additional learning algorithm) can be created.

In this case, the pre-learned algorithm may be an algorithm in which a long short-term memory (LSTM) algorithm that performs time-series monitoring and a random forest algorithm capable of preventing overfitting are fused.

3 is a conceptual diagram for explaining a pre-learned algorithm for calculating a first sleep efficiency index representing objective sleep efficiency and a second sleep efficiency index representing subjective sleep efficiency.

First of all, the LSTM algorithm, as a type of RNN (Recurrent Neural Network), is an excellent algorithm for learning time pattern and time series analysis.

Unlike Deep Feedforward Neural Networks (DFNNs), where neurons are connected in only one direction, RNNs have hidden states that connect the previous state value to the next state value. RNNs with temporal feedback are useful for learning dynamic (time-dependent) mappings from inputs to outputs.

LSTMs are more effective for analyzing time-series data than basic RNNs because they can overcome the vanishing problem and learn long-term dependencies effectively through memory cells and gates.

LSTMs also include a gating mechanism to regulate information flow. Accordingly, the amount of incoming information to be kept is systematically determined at each time step, which allows the LSTM to remember temporal patterns for longer periods of time.

A random forest algorithm refers to a combination of decision tree classifiers in which each decision tree depends on an independently sampled random vector value and has the same distribution over all trees in the forest.

This allows the random forest algorithm to use thousands of variables without overfitting. This algorithm creates an ensemble of trees to improve the limited robustness and performance of decision trees.

Random forest algorithms are widely used in regression and classification problems due to their efficiency and low error rate.

Subsampling data and variables in the random forest algorithm reduces the variance and correlation of tree biases.

In addition, each decision tree of the random forest algorithm is resistant to overfitting because it is learned with various combinations of variables and data, even if there is considerable learning.

Overfitting means overtraining the training data in machine learning.

The pre-learned algorithm of the present invention can be formed to output a sleep efficiency index by combining the advantages of the LSTM algorithm and the random forest algorithm to learn a high-dimensional time-series sleep index without overfitting.

As such, the pre-learned algorithm presented in the present invention may be named LSTM-Forest algorithm or LSTM-Random forest algorithm.

Specifically, the pre-learned algorithm may be formed to use an LSTM algorithm instead of the decision tree in a random forest algorithm using a decision tree.

That is, the present invention can provide an LSTM-Forest algorithm (or model), which is a sleep efficiency index prediction framework that uses LSTM instead of a decision tree in a random forest algorithm.

The LSTM-Forest model can generate multiple LSTM models through random data sampling and random selection of variables.

The LSTM-Forest model is an ensemble model with all trained LSTMs, and each weak LSTM model in this framework uses a small number of variables. However, all variables (information related to sleep) are used in the LSTM-Forest model.

Also, since LSTM-Forest models are not trained with a bias for specific variables and data, LSTM-Forest models can prevent overfitting.

That is, the computer inputs sleep-related information including a plurality of sleep indicators to a pre-learned algorithm (LSTM-Forest model) as an input value, and the first sleep efficiency index (Y1) representing objective sleep efficiency and subjective A second sleep efficiency index Y2 representing sleep efficiency may be output as an output value.

Meanwhile, the pre-learned algorithm of the present invention, as described above, may perform multi-tasking or multi-tasking learning configured to output a plurality of output values.

4 is a conceptual diagram for explaining multi-tasking applied to the pre-learned algorithm of the present invention.

Multitasking means solving multiple learning tasks simultaneously by exploiting similarities and differences between tasks.

Figure 4 (a) shows an LSTM model that performs single-tasking, and Figure 4 (b) shows an LSTM model that performs multi-tasking.

Referring to (b) of FIG. 4, the LSTM model of FIG. 4 simultaneously calculates a first sleep efficiency index (Objective SE) representing objective sleep efficiency and a second sleep efficiency index (Subjective SE) representing subjective sleep efficiency A shared layer and a layer for each task may be formed to do so.

As shown in (b) of FIG. 4, in the LSTM-Forset model (algorithm) of the present invention for performing multi-testing, the shared layer is formed of an LSTM layer and a Full Connected (FC) layer, and The layer for each task that outputs the first and second sleep efficiency indices may be formed as an FC layer.

The computer providing the insomnia treatment method of the present invention may output the first sleep efficiency index and the second sleep efficiency index through multi-tasking applied to the previously learned algorithm. In addition, the computer may also calculate a third sleep efficiency index (Y3) corresponding to the difference between the first sleep efficiency index and the second sleep efficiency index.

In this way, the pre-learned algorithm of the present invention can output a plurality of output values through one algorithm, and thus it is not necessary to have a plurality of algorithms (learning models), and thus overload of data processing can be prevented.

5 is a conceptual diagram for explaining sleep indexes affecting first and second sleep efficiency indices in the present invention.

Meanwhile, as shown in FIG. 5 , the computer of the present invention has a first sleep index (Close, TR, High, Volume, etc.) and a predetermined number of second sleep indexes (Volume, SD20, Close, SRSI10, etc.) related to the second sleep efficiency index may be extracted.

The extracted first sleep index and the second sleep index may refer to a predetermined number of high-order sleep indexes that have a great influence on the first sleep efficiency index and the second sleep efficiency index, respectively.

That is, the pre-learned algorithm provided by the present invention, when outputting the first sleep efficiency index and the second sleep efficiency index, input values that affect the calculation of each sleep efficiency index (i.e., included in sleep-related information) A plurality of sleep indicators) may be calculated. Then, based on the weight, the computer may calculate a predetermined number of high-order sleep indexes having a great influence on the first and second sleep efficiency indices, respectively.

In addition, the computer may also calculate a predetermined number of upper third sleep indices that greatly affect the third sleep efficiency index Y3 corresponding to the difference between the first and second sleep efficiency indices.

Thereafter, the computer may provide a treatment method related to at least one of the first sleep index and the second sleep index.

For example, if the first sleep index that has the most influence on the first sleep efficiency index is drinking coffee before sleep, a treatment method for not drinking coffee from 1 hour before the start of sleep is provided (suggested). can

As another example, in the psychological aspect, if the second sleep index that had the most impact on the second sleep efficiency index was continuous smartphone use, a treatment method was provided to reduce the smartphone use time and exercise one hour before bedtime (suggested )can do.

The computer may provide a treatment method for treatment/intervention to increase at least one of the first and second sleep efficiency indices based on at least one of the first sleep index and the second sleep index.

Treatment methods for treatment/intervention to increase at least one of the first and second sleep efficiency indices include, for example, cognitive correction, stimulation control, sleep restriction therapy, healthy sleep habit education, circadian rhythm control techniques, and stress. management may be included.

Cognitive correction means exploring and correcting cognitive errors about sleep. The present invention, through cognitive correction, can give the patient a general understanding of sleep and insomnia and create a foundation for self-improvement, and make insomnia patients aware of wrong thoughts and habits about sleep and improve sleep quality. It can help you regain the 'power to fall asleep on your own' by putting the contents into practice.

Stimulus conditioning means performing conditioning correction between external stimuli and sleep.

The present invention adjusts the relationship between sleep and stimulation conditions that control sleep, such as bedtime, bedtime, and bedroom, through a stimulus control method, and prohibits reading, watching TV, snacking, etc. in bed or 4-6 hours before going to bed. Corrective actions such as not drinking caffeinated beverages such as coffee and green tea may be suggested.

Sleep restriction therapy refers to an algorithm that optimizes sleep efficiency, and may limit wake-up time.

The present invention, through sleep restriction therapy, prevents the adverse effects on insomnia caused by the behavior of insomnia patients staying in bed for a long time to compensate for insomnia, and by limiting naps, sleep outside the bed, and time in bed, the patient can You can consolidate the sleep you have been having and consequently improve the efficiency of your sleep.

Healthy sleep habits education can be performed through interactive behavioral training.

The circadian rhythm control technique may refer to suggesting environmental illumination for optimizing the sleep cycle.

The present invention, through the circadian rhythm control technique, in order to normalize the circadian rhythm of the 24-hour cycle of the living body, light therapy that exposes artificial light at a certain time of the day or meals at a certain time You can normalize your circadian rhythm by suggesting ways to exercise, work, and so on.

Stress management therapy refers to therapy that relieves anxiety and overindulgence.

The computer may provide at least one of the above-described various treatment methods based on the calculated first and/or second sleep index.

For example, if the treatment method determined based on at least one of the first and second sleep indicators is cognitive correction, the computer may provide an alternative thinking slogan suitable for the individual's cognitive distortion type during cognitive therapy.

For example, computers say, 'Today's sleep affects less than 10% of tomorrow!', 'I suffer from the compulsion to sleep well, not because I can't sleep.' slogans can be provided, and the frequency of presentation can be determined according to the contribution of cognitive distortion among the total treatment factors.

As another example, if the treatment method determined based on at least one of the first and second sleep indexes is stimulation control, the computer reports, 'Of the causes of Mr. OO's insomnia, the habit of lying down before going to bed is the biggest cause at XX%. Even after treatment, please make sure to fix this habit.'

As another example, if the treatment method determined based on at least one of the first and second sleep indexes is sleep restriction therapy, the computer will say, 'The first cause of Mr. OO's insomnia is trying to sleep too much. Please choose what kind of effort you need to make to be able to wake up at the scheduled time even if you go to bed late.' Or 'Napping is a major negative influence on your sleep quality. Try not to nap.'

As another example, if the treatment method determined based on at least one of the first and second sleep indicators is healthy sleep habit education, the computer will say, 'The first cause of OO's insomnia is drinking. For OO, there is a big difference in sleep score between the day he drank and the day he didn't drink, just like OO. Please select one of the plans below to control your drinking in the future.'

As another example, if the treatment method determined based on at least one of the first and second sleep indexes is a circadian rhythm control technique, the computer reports, 'OO's biorhythm is pushed back by 2 hours and 20 minutes. Looking at the light outdoors right after waking up helps to restore your circadian rhythm. You can provide information such as, “Get active during the day.”

As another example, if the treatment method determined based on at least one of the first and second sleep indicators is stress management therapy, the computer may say, 'The first cause of your insomnia is anxiety. Depending on the anxiety index, there was a big difference in sleep scores, such as OO. Watch out for the following ways to control your anxiety.' You can provide comments such as

In addition, the computer may also suggest an insomnia treatment method based on the third sleep index.

Meanwhile, the computer of the present invention may collect information related to new sleep after applying the provided treatment method.

For example, the computer may newly collect information measured using at least one of a mobile terminal and a wearable device and information input by the user the day after the user goes to bed after performing the treatment method.

Then, unlike the previous day, the computer inputs the new sleep-related information and the provided treatment method as input values of the previously learned algorithm to obtain a new first sleep efficiency index and a new second sleep efficiency index as output values. can

Thereafter, the computer determines whether the first sleep efficiency index or the second sleep efficiency index has increased compared to the previous day and whether the difference between the first and second sleep efficiency indexes has decreased, and the insomnia treatment presented through the sleep index of the previous day You can decide whether to keep the method or offer a new treatment for insomnia.

According to the present invention, sleep efficiency can be more accurately measured without using expensive equipment.

In addition, according to the present invention, it is possible to simultaneously calculate objective sleep efficiency and subjective sleep efficiency by inputting high-dimensional sleep-related variables into a time-series and non-overfitting model, thereby obtaining accuracy and speed for measuring sleep efficiency at the same time. can

In addition, according to the present invention, it is possible to provide a more direct and effective treatment method for insomnia by extracting sleep indicators that greatly affect objective sleep efficiency and subjective sleep efficiency and providing a treatment method related thereto.

The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

The method for providing training based on a patient's condition, which is a method according to an embodiment of the present invention described above, is a computer program (or application) for providing training based on a patient's condition to be executed by combining a computer, which is hardware. It can be implemented and stored in a medium.

Steps of a method or algorithm described in connection with an embodiment of the present invention may be implemented directly in hardware, implemented in a software module executed by hardware, or implemented by a combination thereof. A software module may include random access memory (RAM), read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, hard disk, removable disk, CD-ROM, or It may reside in any form of computer readable recording medium well known in the art to which the present invention pertains.

Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art to which the present invention pertains can be implemented in other specific forms without changing the technical spirit or essential features of the present invention. you will be able to understand Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive.

Claims (11)

As a computer device for the treatment of insomnia,
The computer device,
Sleep-related information is input to the pre-learned algorithm as an input value, and a first sleep efficiency index representing objective sleep efficiency and a second sleep efficiency index representing subjective sleep efficiency are obtained as output values,
Based on the first sleep efficiency index and the second sleep efficiency index, providing a treatment method related to sleep,
The pre-learned algorithm,
A computer device for treating insomnia, characterized in that the LSTM algorithm for performing time series modeling and the random forest algorithm capable of preventing overfitting are fused algorithms. According to claim 1,
The information related to the sleep,
A computer device for treating insomnia, characterized in that it includes information measured through at least one of a mobile terminal and a wearable device and input information for a pre-written question. delete According to claim 1,
The pre-learned algorithm,
In a random forest algorithm using a decision tree, a computer device for treating insomnia, characterized in that formed to use an LSTM algorithm instead of the decision tree. According to claim 1,
The pre-learned algorithm,
A computer device for treating insomnia, characterized in that it performs multi-tasking configured to output a plurality of output values. According to claim 5,
The computer device,
Obtaining the first sleep efficiency index and the second sleep efficiency index as the output values through multi-tasking applied to the pre-learned algorithm,
A computer device for treating insomnia, characterized in that a third sleep efficiency index corresponding to a difference between the first sleep efficiency index and the second sleep efficiency index is calculated and further obtained as the output value. According to claim 1,
The information related to sleep includes a plurality of sleep indicators,
A computer device for treating insomnia, characterized in that a first sleep index related to the first sleep efficiency index and a second sleep index related to the second sleep efficiency index are extracted by a predetermined number among the plurality of sleep indexes. . According to claim 7,
For the treatment of insomnia, characterized in that the extracted first sleep index and the second sleep index are the upper predetermined number of sleep indexes that have a great influence on the first sleep efficiency index and the second sleep efficiency index, respectively. computer device. According to claim 8,
The treatment method is characterized in that related to at least one of the first sleep index and the second sleep index, a computer device for treating insomnia. According to claim 1,
The computer device,
Collecting information related to new sleep after applying the provided treatment method,
A computer device for treating insomnia, which obtains a new first sleep efficiency index and a new second sleep efficiency index as output values by inputting the new sleep-related information and the provided treatment method as input values of the pre-learned algorithm. delete

Images