KR102053604B1 - Method for sleeping analysis and device for sleeping analysis using the same - Google Patents

Abstract

In the present specification, a sleep evaluation model configured to receive non-sleep state data and sleep state data from a wearable device and evaluate an influencing factor of sleep based on the non-sleep state data and the sleep state data for the user, Provided are a sleep analysis method and a sleep analysis device using the same, which are configured to evaluate the influencing factors of sleep.

Description

Sleep analysis method and sleep analysis device using the same {METHOD FOR SLEEPING ANALYSIS AND DEVICE FOR SLEEPING ANALYSIS USING THE SAME}

The present invention relates to a sleep analysis method and a sleep analysis device using the same. More particularly, the present invention relates to a sleep analysis method and a sleep analysis device using the same for evaluating factors affecting sleep and predicting sleep quality.

Sleep accounts for about one third of human life and plays important roles in energy storage and recovery, as well as hormone secretion and memory coagulation. Lack of sleep or poor sleep can lead to blurred mental activity and insensitivity. In modern society, more and more people are complaining of sleep disorders due to environmental factors such as aging population, increased stress, aging and changes in sleep cycle.

Furthermore, sleep has been reported to be associated with mental illnesses such as depression. For example, if you have or have a severe mental illness, your sleep may be worse. In addition, when the degree of mental illness improves, an improvement in the sleep state may appear as an indicator. As such, the sleep state may be used as important information regarding the extent, prognosis, or therapeutic effect of mental illness.

Thus, in the treatment of sleep disorders, it may be an important factor to accurately analyze the sleep state for each individual and to understand what factors affect sleep.

On the other hand, sleep disorders occur during sleep, so it is difficult for you to identify symptoms. In addition, polysomnography, which comprehensively examines EEG, muscle movement, respiration, electrocardiogram, and oxygen saturation during sleep, can be used to accurately diagnose sleep disorders, but due to the high cost of analytical testing and treatment Is a difficult situation. Furthermore, such sleep analysis methods are not suitable to be performed at home due to the uncertainty due to errors, and a large number of electrodes must be attached to the user's head.

Accordingly, the device-based sleep analysis method can put a heavy burden on the patient. Furthermore, these factors can be perceived by many patients as unstable experience in sleep analysis, and can affect the results of the analysis.

Accordingly, it is required to develop a system that can overcome the above limitations and analyze the quality of sleep effectively. In particular, development of a sleep analysis method and device for analyzing sleep quality of a user having a sleep disorder, analyzing the cause thereof, and providing the user to the user may be very important in preventing, diagnosing, and treating sleep disorders. have.

The background art of the invention has been written in order to facilitate understanding of the present invention. It should not be understood that the matters described in the background of the invention exist as prior art.
* Prior art document number: Korean Patent Publication No. 10-2017-0089231

The wearable device is used in the form of a smart band, a smart watch, a wearable glass, and the like, and detects a biorhythm using a sensor provided in the wearable device, and provides a service using the wearable device.

As an example of using the wearable device, it is possible to determine whether the user has had a good sleep time by analyzing a sleep pattern by detecting a sleep state of the user. In this case, the wearable device measures information on a user's body temperature, pulse rate, etc. using a sensor provided therein, and analyzes the sleep state of the user based on the information.

However, conventional wearable devices have a problem in that the types and amounts of data that can be measured are limited, and the functions that can be performed are also limited. In addition, in order to measure the sleep state, there is an inconvenience in that a user directly inputs a control command for starting and ending the mode of the wearable device when sleep starts and ends, and performs such an input process immediately before and after sleep. Because the user has to immerse the sleep, the user may not have a good night's sleep. And, when analyzing the sleep state using only the data measured in the wearable device, the room for error increases. For example, in a situation such as meditating or reading a book using only one arm partially, the awakening state may be mistaken as a sleep state.

Accordingly, a sleep analysis system using a wearable device may require a new technology for predicting a sleep quality of a user and in particular, evaluating sleep disturbance factors of a user having a sleep disorder.

Meanwhile, the inventors of the present invention pay attention to activities in the non-sleep state, the surrounding environment, and anxiety with respect to sleep disorders, and have recognized that these factors are associated with sleep disorders.

Furthermore, the inventors of the present invention recognized that the evaluation model learned by the state data associated with the quality of sleep of the user can be used to improve the accuracy of the prediction about the quality of sleep and to analyze the obstacles of the sleep.

As a result, the inventors of the present invention can analyze the sleep disturbance factors and predict the sleep quality based on the data such as the amount of activity, the intensity of illumination, the change of the pulse rate, the state and the mode change, etc. To develop a new sleep analysis system.

Accordingly, an object of the present invention is to receive data in a non-sleep state from a wearable device from a user, predict sleep quality based on the sleep quality prediction model and the sleep evaluation model, and sleep disorders The present invention provides a sleep analysis method for evaluating factors and a sleep analysis device using the same.

The problem to be solved by the present invention is to directly receive the sleep record data associated with sleep at a predetermined time period from the user, to predict the sleep quality based on the sleep record data using the sleep quality prediction model and the sleep evaluation model The present invention provides a sleep analysis method for evaluating sleep disorders and a sleep analysis device using the same.

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

In order to solve the above problems, there is provided a sleep analysis method according to an embodiment of the present invention. The sleep analysis method of the present invention includes a sleep evaluation model configured to evaluate non-sleep state data and sleep state data from a wearable device for a user, and evaluate influence factors of sleep based on the non-sleep state data and the sleep state data. Using, evaluating factors affecting sleep on the user.

According to a feature of the invention, the step of receiving may include receiving each of the non-sleep state data and the sleep state data in a plurality of time quantum, and the step of evaluating the factors affecting sleep using a sleep evaluation model The method may include evaluating factors affecting sleep according to the time quartile based on the non-sleep state data and the sleep state data received in the plurality of time quartiles.

According to another feature of the invention, the step of evaluating the influencing factors of sleep may further include calculating a relative or absolute degree of sleep influence on the influencing factors of sleep over time.

According to another aspect of the present invention, the sleep analysis method comprises the steps of generating a sleep influence configured to have an area size proportional to the degree of sleep influence corresponding to the influence factors of sleep over time quartile, and to provide sleep influence It may further comprise the step.

According to another aspect of the invention, the sleep analysis method, generating a sleep impact distribution map configured to have a distribution degree proportional to the sleep influence degree corresponding to the influence factors of sleep over time quartile, and provides a sleep impact distribution map It may further comprise the step.

According to another feature of the present invention, the receiving may include receiving non-sleep state data and sleep state data at a plurality of time quartiles, for each of the plurality of dates. Further, the step of evaluating the influence factors of the sleep according to the time quartile is a time quartile based on the non-sleep state data and the sleep state data received in the plurality of time quartiles for each of the plurality of days using the sleep evaluation model. It may include the step of evaluating the factors affecting the sleep according to. The sleep analysis method may further include receiving a specific date selected from a plurality of dates, and providing an influence factor of sleep according to a time quartile corresponding to the specific date.

According to another feature of the present invention, the receiving may include receiving, from the wearable device, non-sleep state data and sleep state data for a predetermined period of time. Furthermore, the sleep analysis method may further include generating a time series state analysis diagram for each of the non-sleep state data and the sleep state data for a predetermined period, and providing the time series state analysis diagram.

According to another feature of the invention, the sleep analysis method may further include extracting a feature variable for each of the non-sleep state data and the sleep state data, which is performed after the receiving step. Furthermore, the step of evaluating the influence factor of sleep may further include evaluating the influence factor of sleep on the user using the sleep evaluation model based on the characteristic variables of each of the non-sleep state data and the sleep state data. Can be.

According to another aspect of the present invention, the sleep analysis method, using the sleep quality prediction model configured to predict sleep quality based on the non-sleep state data, predicting sleep quality of the user and predicted sleep The method may further include providing a plurality of levels of quality.

According to another feature of the present invention, the step of evaluating the influence factor of the sleep may further include calculating a sleep efficiency of the user based on the sleep state data.

According to another feature of the present invention, the non-sleep state data may be at least one selected from the group consisting of activity amount, illuminance amount, heart rate, and HRV measured by the wearable device in the user's non-sleep state.

According to another feature of the present invention, the sleep state data may include sleep measurement data measured in the sleep state of the user by the wearable device and sleep record data input at predetermined intervals from the user. In this case, the measurement data during sleep may be at least one of a group consisting of HRV, heart rate, illuminance, and sleep activity measured in the user's sleep state. Furthermore, the sleep history data may include the date, name, gender, age, time of day, time of sleep, day of sleep, quality of sleep on the day, degree of freshness on the day, no-wear period of the wearable device, caffeine intake, alcohol intake, nap or not. And at least one selected from BMI.

In order to solve the above problems, there is provided a sleep analysis device according to an embodiment of the present invention. The sleep analysis device of the present invention includes a receiver configured to receive non-sleep state data and sleep state data for a user from a wearable device, and a processor coupled with the receiver. In this case, the processor is configured to evaluate the influence factor of sleep on the user using a sleep evaluation model configured to evaluate the influence factor of sleep based on the non-sleep state data and the sleep state data.

According to a feature of the invention, the receiving unit is configured to receive the non-sleep state data and the sleep state data, respectively, in a plurality of time quartiles, and the processor uses the sleep evaluation model to receive the non-sleep state data received in the plurality of time quartiles It may be further configured to evaluate the influencing factor of sleep over time based on the sleep state data.

According to another feature of the invention, the processor may be further configured to calculate a relative or absolute degree of sleep influence on the factors influencing sleep over time quartile.

According to another feature of the invention, the processor is further configured to generate a sleep influence configured to have an area size proportional to the sleep influence degree corresponding to the sleep influence factor over time and to provide a sleep influence. Can be.

According to another feature of the invention, the processor may be further configured to generate a sleep influence distribution plot configured to have a distribution degree proportional to a sleep influence degree corresponding to a sleep influence factor over time, and to provide a sleep influence distribution map. Can be.

According to another feature of the present invention, the receiver may be further configured to receive, for each of the plurality of dates, non-sleep state data and sleep state data, and a specific date selected from the plurality of dates at a plurality of time quartiles. The processor uses the sleep assessment model to evaluate the influence factors of sleep over time quartiles based on the non-sleep state data and the sleep state data received at the plurality of time quartiles for each of the plurality of days, and a specific date. It may be further configured to provide an influencing factor of sleep according to the time quartile corresponding to the.

According to another feature of the invention, the receiver may be further configured to receive the non-sleep state data and the sleep state data for the predetermined period from the wearable device. Further, the processor may be further configured to generate a time series state analysis diagram for each of the non-sleep state data and the sleep state data for the predetermined period of time, and provide the time series state analysis diagram.

According to another feature of the invention, the processor predicts the sleep quality of the user by using a sleep quality prediction model configured to predict the sleep quality based on the non-sleep state data, and calculates the predicted sleep quality It may be further configured to provide at a plurality of levels.

According to another feature of the invention, the processor may be further configured to calculate a sleep efficiency of the user based on the sleep state data.

According to another feature of the invention, the processor may be further configured to extract feature variables for each of the non-sleep state data and the sleep state data. Furthermore, the sleep evaluation model may be further configured to evaluate the factors influencing sleep on the user based on the feature variables for each of the non-sleep state data and the sleep state data.

Hereinafter, the present invention will be described in more detail with reference to Examples. However, these examples are only for illustrating the present invention by way of example, and the scope of the present invention should not be construed as being limited by these examples.

The present invention provides a sleep analysis method using a sleep evaluation model configured to evaluate sleep based on biometric data measured in a non-sleep state that may be related to sleep quality, and a sleep quality prediction model configured to predict sleep, and using the same. Providing a sleep analysis device may enable accurate cause analysis of a user's sleep disorder. As a result, the present invention has the effect of providing a personalized prescription according to the analysis results.

That is, the present invention may provide and evaluate factors affecting the user's sleep based on non-sleep state data measured from the wearable device, more specifically, sleep variables during the day, and predict and provide a quality of sleep. can do.

Thus, the present invention has the effect that can solve problems such as expensive analysis cost, inconvenience that the user has a conventional sleep analysis system.

The present invention further predicts sleep quality by considering factors such as sleep history data input directly from a user, together with non-sleep state data, such as caffeine intake, alcohol intake, time to sleep, and degree of freshness. For a plurality of users having a pattern, it is possible to evaluate factors affecting sleep and predict sleep quality with high accuracy and sensitivity.

The present invention can calculate the sleep efficiency of the user based on the biometric data measured in the sleep state and provides a sleep analysis method using a sleep evaluation model and a sleep analysis device using the same, if the quality of sleep is poor There is an effect that can provide cause and further diagnostic feedback.

More specifically, the present invention calculates sleep efficiency using biosignal data provided by a wearable device and sleeps associated with sleep quality to a user using a sleep evaluation model trained on biometric data related to sleep quality. It has the effect of providing analytical information.

Thus, the present invention has the effect of recognizing the cause of insomnia for a user with a sleep disorder, and providing information to improve it.

The effects according to the present invention are not limited by the contents exemplified above, and more various effects are included in the present specification.

1 illustrates a sleep analysis method and a sleep analysis device based sleep analysis system using the same according to an embodiment of the present invention.
2 exemplarily shows a configuration of a sleep analysis device according to an embodiment of the present invention.
3 and 4 exemplarily illustrate a procedure for a sleep analysis method according to various embodiments of the present disclosure.
5A is a diagram illustrating a screen on which state data of a user, received according to a sleep analysis method according to an embodiment of the present invention, is shown.
5B is a diagram exemplarily illustrating a screen on which a time series state analysis diagram of a user is generated according to a sleep analysis method according to an exemplary embodiment of the present invention.
5C is a diagram exemplarily illustrating a screen showing sleep influence factors and predicted sleep quality evaluated according to a sleep analysis method according to an exemplary embodiment of the present invention.
FIG. 6A illustrates non-sleep state data obtained from a wearable device used in a sleep analysis method and a sleep analysis system using the same according to an embodiment of the present invention.
FIG. 6B illustrates characteristic variables of non-sleep state data and sleep state data obtained from a wearable device, which are used in a sleep analysis method and a sleep analysis device using the same according to an embodiment of the present invention.
FIG. 6C illustrates sleep recording data used in a sleep analysis method and a sleep analysis device using the same according to an embodiment of the present invention.
7A to 7C illustrate evaluation results of sleep disturbance factors based on a sleep evaluation model analyzed in a sleep analysis method and a sleep analysis system using the same according to an embodiment of the present invention.
7D illustrates a sleep quality prediction result based on a sleep quality prediction model analyzed in a sleep analysis method and a sleep analysis system using the same according to an embodiment of the present invention.

Advantages and features of the present invention, and methods for achieving them will become apparent with reference to the embodiments described below in detail 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 the present embodiments to make the disclosure of the present invention complete, and common knowledge in the art to which the present invention pertains. It is provided to fully inform the person having the scope of the invention, which is defined only by the scope of the claims.

Hereinafter, a sleep analysis method and a sleep analysis device based on the sleep analysis device using the same according to an embodiment of the present invention will be described in detail with reference to FIGS. 1 to 4.

1 illustrates a sleep analysis method and a sleep analysis device based sleep analysis system using the same according to an embodiment of the present invention. 2 exemplarily shows a configuration of a sleep analysis device according to an embodiment of the present invention. 3 and 4 exemplarily illustrate a procedure for a sleep analysis method according to various embodiments of the present disclosure.

First, referring to FIG. 1, the sleep analysis system 1000 measures sleep state data or sleep state data for a user, a sleep analysis device 100, or records sleep from a user according to an embodiment of the present invention. Wearable device 200 configured to receive data, bed stationary device 300 configured to measure sleep state data of a user, and clinician device 400 configured to receive sleep influence factors evaluated by the user and predicted sleep quality It can be configured as.

According to the sleep analysis system 1000 of the present invention, the sleep analysis device 100 according to an embodiment of the present invention is a data of the amount of activity, illuminance, heart rate and HRV in the non-sleep state of the user from the wearable device 200. Non-sleep state data, and / or HRV levels, heart rate, and sleep activity in the sleep state, and based on the assessment of sleep influencing factors or predict sleep quality may be provided to the medical staff device 400. .

In this case, the sleep state data may be measured from the bed fixed device 300, received by the wearable device 200, and then received by the sleep analysis device 100, and the sleep analysis device 100 may be received from the bed fixed device 300. May be directly received).

The sleep analysis device 100 may be configured to use a sleep evaluation model configured to evaluate a factor influencing sleep based on state data of the user and a sleep quality prediction model configured to predict sleep quality. Meanwhile, the sleep evaluation model and the sleep quality prediction model are based on the correlation between the variables such as acceleration, illuminance, pressure, HR, HRV, time, and events in the non-sleep state acquired in advance and sleep retardation (sleep efficiency). It can be based on a learned prediction model or prediction algorithm. At this time, each model may be used singly or in combination.

Meanwhile, the wearable device 200 may include an acceleration sensor, an illuminance sensor, a PPG sensor, an event button, a clock, and the like, and collect events in major non-sleep states such as activity amount, brightness, pulse rate, time zone, and bedtime. Can be transmitted to the sleep analysis device 100. Furthermore, the wearable device 200 may include a date, a name, a gender, an age, a time that happened on the day, a day before sleep, a quality of sleep on the day before, the degree of freshness on the day, a non-wearing period of the wearable device, a caffeine intake, an alcohol intake, a nap And receive sleep recording data such as BMI.

The sleep-fixed device 300 may include a pressure sensor or the like to measure the amount of activity of the user during sleep, that is, the degree of twisting. Sleep state data measured by the sleep stationary device 300 may be received at the sleep analysis device 100 after being received at the wearable device 200 as described above, and may be received directly at the sleep analysis device 100. It may be.

Sleep influence factors and predicted sleep quality for the user evaluated by the sleep analysis device 100 may be transmitted to the wearable device 200 worn by the user so that sleep information may be provided to the user. Further, sleep influencing factors and predicted sleep quality may be transmitted to the clinician device 400 so that the clinician can more easily take appropriate medical measures for the user based on the evaluated and predicted results for the user.

Thus, the sleep analysis device 100 of the present invention may be applied to the sleep analysis system 1000 as it may provide a sleep analysis result to both the medical staff and the user.

 Referring to FIG. 2, the sleep analysis device 100 may include a receiver 110, an inputter 120, a display 130, a storage 140, and a processor 150.

More specifically, the receiver 110 may include one or more modules that enable wired and wireless communication between the device and the network in which the device is located. The receiver 110 transmits and receives a signal with an external device, for example, a wearable device, on a communication network such as the Internet. More specifically, the receiver 110 may receive the activity amount, the illuminance amount, the heart rate, and the non-sleep state data of the HRV measured in the non-sleep state from the wearable device of the user. Furthermore, the receiver 110 may measure the HRV level in the sleep state measured from the wearable device in the sleep state of the user, the heart rate in the sleep state, and the sleep activity amount measured from the bed-fixed device (return during sleep) during sleep measurement data and / Or the date, name, gender, age, the time of day, the day before sleep, the quality of sleep the day before, the degree of freshness on the day, the wearing period of the wearable device, caffeine intake, alcohol intake, nap and sleep history data of BMI Can be received.

The input unit 120 generates input data for the user to control the operation of the device. In this case, the input unit 120 may include a key pad dome switch, a touch pad (constant voltage / capacitance), a jog wheel, a jog switch, and the like. However, it is not limited thereto. On the other hand, the user through the input unit 120, the date, name, gender, age, the time of the day, the day before sleeping, the quality of sleep the day before, the degree of freshness, the wearing period of the wearable device, caffeine intake alcohol intake, nap or not And sleep recording data of the BMI can be directly input, but is not limited thereto.

The display unit 130 is for generating an output related to sight, hearing, or tactile sense, and may include a display unit, an audio output module, and the like, wherein the display unit is configured to display (output) information processed by the device. Can be. For example, the display unit 130 displays a user interface (UI) or graphic user interface (GUI) in which the device is associated with the system. In this case, the display unit 130 may include a liquid crystal display (LCD), a thin film transistor-liquid crystal display (TFT LCD), an organic light-emitting diode (OLED), a flexible display ( It may include at least one of a flexible display, a 3D display. The display 130 may be further configured to output audio data received from the receiver 110 or stored in the storage 140 to be described later. In this case, the display unit 130 may output a sound signal related to a function performed by the sleep analysis device 100. In various embodiments, the input unit 120 or the display unit 130 may be implemented to be omitted.

The storage unit 140 may store a program for processing and controlling the processor 150, which will be described later, or perform a function for temporarily storing input / output data. For example, the storage 140 may be configured to store non-sleep state data and / or sleep state data received by the receiver 110.

Meanwhile, the storage 140 may include a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (eg, SD or XD memory, etc.). Random Access Memory (RAM), Static Random Access Memory (SRAM), Read-Only Memory (ROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Programmable Read-Only Memory (PROM), Magnetic At least one type of storage medium may include a memory, a magnetic disk, and an optical disk. In this case, the sleep analysis device 100 may operate in association with a web storage that performs a storage function of the storage 140 on the Internet.

The processor 150 can typically control the overall operation of the sleep analysis device 100. For example, the processor 150 inputs non-sleep state data and / or sleep state data for the user received by the receiver 110 into the sleep evaluation model and the sleep quality prediction model, respectively, and the user uses these prediction models. It may be configured to assess sleep disorders and predict sleep quality according to the non-sleep and / or sleep state of.

According to an embodiment of the present invention, the processor 150 may extract feature variables for a plurality of quartiles from the non-sleep state data and / or the sleep state data for the user obtained from the wearable device, and use the sleep evaluation model. And assess sleep disturbances based on the extracted feature variables. For example, the processor 150 may display data of the amount of activity for the user received by the receiver 110 in the first quartile from the wake-up time to 12 pm, the second quartile from 12 pm to 6 pm, and the six pm portion. In each of the quartiles from the beginning of the day to bedtime, the characteristics can be classified by intensity of upper, middle and lower on the basis of the Metabolic Equivalent of Task (MET). Further, the processor 150 may evaluate the sleep disorder factor based on the feature variable for the amount of activity using the sleep evaluation model.

According to another embodiment of the present invention, the processor 150 predicts sleep quality for the user by using the sleep quality prediction model based on the non-sleep state data of the user received by the receiver 110. It can be further configured to. For example, the processor 150 may predict the quality of sleep of the user as good or bad based on the activity amount, illuminance amount, heart rate, and measurement data by the wearable device of the HRV for the user received by the receiver 110. .

Meanwhile, the processor 150 may further include a graphic module for parallel data processing, which may be implemented in the processor 150 or may be implemented separately from the processor 150.

The various embodiments described herein may be implemented in a recording medium readable by a computer or similar device using, for example, software, hardware or a combination thereof.

According to a hardware implementation, the embodiments described herein include application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), and the like. It may be implemented using at least one of processors, controllers, micro-controllers, microprocessors, and electrical units for performing other functions. The described embodiments can be implemented in the processor 150 itself.

According to the software implementation, embodiments such as the procedures and functions described herein may be implemented as separate software modules. Each of the software modules may perform one or more functions and operations described herein. Software code may be implemented in software applications written in a suitable programming language. The software code is stored in the storage 140 and can be executed by the processor 150.

The sleep analysis device 100 of the present invention having the above configuration provides a user with a sleep disorder or a user who wants to predict and evaluate the quality of his sleep in a wide range of information on the quality of his sleep. Can be. For example, a user with a sleep disorder may be provided with information about a habit in a non-sleep state that may be a cause of a sleep disorder through the sleep analysis device 100. In addition, the user may be provided with feedback from the medical staff regarding the sleep quality evaluation result or the sleep quality prediction result.

Hereinafter, a sleep analysis method according to various embodiments of the present disclosure will be described with reference to FIGS. 3 and 4.

First, referring to FIG. 3, the sleep analysis method according to an exemplary embodiment of the present disclosure receives sleep state data and non-sleep state data for a user from a wearable device worn by the user (S310), and sleep state data and non-sleep. A feature variable for the sleep state data is extracted (S320), and an influence factor of sleep is evaluated based on the received feature variables using the sleep evaluation model (S330), and an evaluation result is provided (S340).

More specifically, in the step of receiving (S310), the amount of activity, illuminance, heart rate and HRV level measured from the user's wearable device may be received.

Further, in the receiving step (S310), the sleep time, such as HRV level of the sleep state measured from the wearable device in the user's sleep state, heart rate of the sleep state, and the amount of sleep activity received from the bed fixed device (turns back during sleep) Measurement data and / or date entered by the user, name, gender, age, time of day, time of day of sleep, day of sleep, quality of sleep the day before, day wear of the wearable device, caffeine intake, alcohol intake, nap and BMI Can receive sleep recording data.

According to another feature of the invention, in the receiving step (S310), it is possible to receive each of the non-sleep state data and the sleep state data in a plurality of time quartiles. In this case, the plurality of time quartiles may mean time intervals that are regularly divided for 24 days. For example, in the step S310 of receiving, the non-sleep state data is the first quartile from wake-up time to 12 pm, the second quartile from 12 pm to 6 pm, and the third quartile from 6 pm to bedtime, respectively. May be received at.

According to another feature of the invention, in the receiving step (S310), the non-sleep state data and the sleep state data may be received for a plurality of days within a predetermined period of time at a plurality of time quartiles.

Meanwhile, in the receiving step S310, non-sleep state data and sleep state data may be received for a random continuous time (eg, 24 hours). Further, the non-sleep state data and the sleep state data received during the continuous time may be divided into predetermined time quartiles.

Next, in step S320 of extracting feature variables for sleep state data and non-sleep state data, feature values for non-sleep state data and / or sleep state data obtained from the wearable device may be extracted.

For example, in the step S320 of extracting the feature variable, data of the amount of activity in the non-sleep state is included in the first quartile from the wake-up time to 12 pm, the second quartile from 12 pm to 6 pm, and from 6 pm In each of the quartiles up to the time before bed, the characteristics can be classified by intensity of upper, middle and lower on the basis of MET. Further, in the step of extracting the feature variable (S320), the characteristics may be classified into the intensity of the indoor low light, the indoor high light and the outdoor low light, and the outdoor high light on the basis of the LUX value of the light intensity measured in the non-sleep state. Further, in the step S320 of extracting the feature variable, the HRV value measured in the non-sleep state may include a standard deviation of normal to normal heart rate (SDNN), a root mean square successive difference (RMSD) representing a degree of parasympathetic nerves, PNN50, which indicates the degree of sympathetic nerve, and average heart rate.

Next, in step S330 of evaluating the influence factor of the sleep, the influence factor of the sleep may be predicted based on the feature variables for the non-sleep state data and / or the sleep state data using the sleep evaluation model. For example, in the step of evaluating the influence factor of the sleep (S330), if the feature variables for the non-sleep state data and / or the sleep state data are input to the sleep evaluation model, main features affecting the sleep may be extracted. have.

ive Bayes), LNB (locally weighted na

ive Bay), 및 SVM (support vector machine) 일 수 있으나 이에 제한되는 것은 아니며, 보다 다양한 분류/회귀분석 알고리즘에 기초할 수 있다. The sleep assessment model can then be based on an algorithm configured to classify certain classes from multiple decision trees for classification or to output mean predictions for regression analysis. The sleep evaluation model disclosed in the present invention is random forest (random forest), GNB (Gaussian na)

ive Bayes), LNB (locally weighted na

ive Bay, and support vector machine (SVM), but is not limited thereto, and may be based on more various classification / regression algorithms.

According to a feature of the present invention, in the step of evaluating the influence factor of the sleep (S330), the relative or absolute sleep influence degree on the influence factor of the sleep according to the time quartile can be calculated.

According to another feature of the invention, in the step (S330) of evaluating the influence factors of the sleep, using the sleep evaluation model, the non-sleep state data and the sleep state received at a plurality of time quartiles for each of the plurality of dates, respectively. Based on the feature variables of the data, factors influencing sleep over time can be predicted.

According to another feature of the present invention, in the step of evaluating the influence factor of the sleep (S330), the sleep efficiency of the user may be calculated based on the sleep state data.

In this case, the sleep efficiency may be calculated by Equation 1 below.

 [Equation 1]

) * 100 Sleep efficiency (%) = (1-

) * 100

는 수면 중 활동량을 의미할 수 있고, OBT- IBT는 잠을 청한 시간 (in bed time) 내지 기상 시간의 구간을 의미할 수 있고,

는 각성 상태 (wakefulness) 의 총합을 의미할 수 있다.here,

May refer to the amount of activity during sleep, OBT-IBT may refer to the interval between bed time and wake-up time,

May mean the sum of wakefulness states.

According to a feature of the present invention, in the sleep analysis method of the present invention, the area size proportional to the degree of sleep influence corresponding to the influence factor of sleep according to the time quartile obtained in the step (S330) of evaluating the above-described influence factors of sleep The step of generating a sleep impact configured to have may be further performed. Furthermore, in the sleep analysis method of the present invention, the step of generating a sleep influence distribution map configured to have a distribution degree proportional to the sleep influence degree corresponding to the influence factors of sleep according to the time quartile may be further performed.

Thus, in the providing step (S340), the generated sleep impact or sleep impact distribution can be provided.

According to another aspect of the present invention, in the sleep analysis method of the present invention, the step of generating a time series state analysis chart for the non-sleep state data and the sleep state data for the predetermined period of time received in step (S310) further Can be performed.

Thus, in the providing step (S340), the generated time series state analysis diagram may be provided.

According to the above-described sleep analysis method according to an embodiment of the present invention, the user may be provided with information on sleep influencing factors evaluated to affect sleep. Furthermore, the medical staff may facilitate sleep monitoring of the user, and may more easily take medical measures according to sleep influencing factors.

Hereinafter, a sleep analysis method according to another exemplary embodiment of the present invention will be described with reference to FIG. 4.

First, referring to FIG. 4, a sleep analysis method according to another embodiment of the present invention receives non-sleep state data of a user from a wearable device worn by a user (S410), and receives the data using a sleep quality prediction model. The sleep quality is predicted based on the non-sleep state data during the day (S420), and the prediction result is provided (S430).

More specifically, in the step of receiving (S410), the activity amount, illuminance amount, heart rate and HRV level measured in the non-sleep state can be received from the wearable device of the user.

According to a feature of the invention, in the receiving step (S410), it is possible to receive the non-sleep state data in a plurality of time quartiles. In this case, the plurality of time quartiles may mean time intervals that are regularly divided for 24 days. For example, in the step S410 of receiving, the non-sleep state data is the first quartile from the wake-up time to 12 pm, the second quartile from 12 pm to 6 pm, and the third quartile from 6 pm to bedtime, respectively. May be received at. However, the present invention is not limited thereto, and the time quartile may be set to have a wider range.

 According to a feature of the present invention, in the receiving step (S410), it is possible to receive the non-sleep state data for a plurality of days within a predetermined period, a plurality of time quartiles.

Next, in step S420 of predicting sleep quality, the sleep quality may be predicted based on the received non-sleep state using the sleep quality prediction model.

In this case, the sleep quality prediction model may be based on a deep learning algorithm configured to learn non-sleep state data based on an artificial neural network. For example, the sleep quality prediction model of the present invention may include a deep neural network (DNN), a convolutional neural network (CNN), a deep convolution neural network (DNN), a recurrent neural network (RNN), a restricted boltzmann machine (RBM), The prediction model may be based on a deep belief network (DBN) or a single shot detector (SSD) model, but is not limited thereto.

On the other hand, the sleep quality predicted in the sleep analysis method according to another embodiment of the present invention may be evaluated by the sleep efficiency calculated based on the sleep state data received in the receiving step (S410). For example, the sleep quality predicted based on the non-sleep state data before going to bed by the sleep quality prediction model may be evaluated by comparing the sleep efficiency calculated based on the sleep state data after waking up.

In this case, the sleep efficiency may be calculated by Equation 1 below.

 [Equation 1]

) * 100 Sleep efficiency (%) = (1-

) * 100

는 수면 중 활동량을 의미할 수 있고, OBT- IBT는 잠을 청한 시간 (in bed time) 내지 기상 시간의 구간을 의미할 수 있고,

는 각성 상태 (wakefulness) 의 총합을 의미할 수 있다.here,

May refer to the amount of activity during sleep, OBT-IBT may refer to the interval between bed time and wake-up time,

May mean the sum of wakefulness states.

Therefore, the sleep quality prediction model is applied to the predicted prediction model or the prediction algorithm based on the correlation between the previously acquired non-sleep state data (variable) and the sleep efficiency calculated by the sleep state data. Can be based.

In the providing step S430, the sleep quality predicted in the step S420 of predicting sleep quality may be provided at a plurality of levels, for example, good or bad, upper, middle or lower. However, it is not limited thereto.

According to the sleep analysis method according to another embodiment of the present invention, the user may be provided with information on the quality of sleep predicted based on the non-sleep state data such as the amount of activity during the day or environmental factors.

Hereinafter, a sleep analysis method according to various embodiments of the present disclosure will be described with reference to FIGS. 5A to 5C by taking the screen of the sleep analysis device of the present invention displayed and displayed with the evaluated and predicted sleep analysis information as an example. However, the sleep analysis method of the present invention is not limited to the following, and may be configured to display more various information associated with sleep analysis.

5A is a diagram illustrating a screen on which state data of a user, received according to a sleep analysis method according to an embodiment of the present invention, is shown. 5B is a diagram exemplarily illustrating a screen on which a time series state analysis diagram of a user is generated according to a sleep analysis method according to an exemplary embodiment of the present invention. 5C is a diagram exemplarily illustrating a screen showing sleep influence factors and predicted sleep quality evaluated according to a sleep analysis method according to an exemplary embodiment of the present invention.

Referring to FIG. 5A, a menu field 132 and a data management column 134 are displayed on the display unit 130 of the sleep analysis device 100 of the present invention. More specifically, the menu field 132 displays user selectable patient management, data management, analysis, setting and control (device) menus. Here, when the data management menu is selected by the user, the data management column 134 is displayed on the display unit 130. In this case, the data management column 134 may display data received from the wearable device worn by the examinee. More specifically, in the data management column 134, the data such as the date of receiving the data, the file number, the subject name (patient name), the subject management number (patient management number), the wearable device name (equipment name), the file name, etc. Status data received from the wearable device may be displayed. Such state data may be used as input data for the predictive model of the sleep analysis device of the present invention.

Referring to FIG. 5B, when the analysis menu is selected in the menu field 132 by the user, the time series sleep analysis field 136 is displayed on the display unit 130. In this case, the time series sleep analysis column 136 shows a time series state in which the non-sleep state and the sleep state data (eg, HR, LUX, METs) received from the wearable device of the examinee are analyzed in time series for the period selected by the user. Analysis diagram 136 (a) is shown. Further, the time series sleep analysis column 136 is further shown a section selection bar 136 (b) configured to select a desired period of time. Thus, the status data of the subject analyzed in time series may be provided.

Referring to FIG. 5C, when the analysis menu is selected in the menu field 132 by the user, the sleep distribution analysis field 138 is displayed on the display unit 130. In this case, in the sleep distribution analysis column 138, non-sleep state data of the subject by the sleep quality model and the sleep evaluation model predicted on the basis of the non-sleep state data of the subject by the sleep prediction model for a specific period selected by the user and / Or sleep influencers evaluated based on sleep state data. More specifically, the sleep distribution analysis column 138 includes a sleep quality prediction distribution map (138 (a)), a sleep impact map (138 (b)), a sleep impact map (138 (c)), and a sleep quality index (138 (d). )) Appears.

More specifically, in the sleep quality prediction distribution map 138 (a), the sleep quality predicted by the sleep quality prediction model for the corresponding date, for example, January 3, 2018, is' good 'or' May appear as bad. Sleep Influence (138 (b)) includes factors that influence sleep, determined by the sleep assessment model for a date, such as January 3, 2018. In this case, the sleep influence degree (138 (b)) may appear to have an area in proportion to the sleep influence factor, that is, an area proportional to the sleep influence degree. Accordingly, the user can easily identify the factors having a high degree of sleep influence by looking at the size of the area. On the other hand, the user may select a specific date of the plurality of dates, to check the sleep impact (138 (b)) for the date. The sleep influence distribution map 138 (c) may be configured to have a distribution degree proportional to the sleep influence degree for a plurality of time quartiles determined by the sleep evaluation model, for example, the sleep influence factors according to the 1st to 3rd quartiles. have. Accordingly, the user can easily identify the time quartile and the sleep influencing factor having a great influence on the sleep by comparing the distribution degree. Meanwhile, the user may select a specific date from among the plurality of dates, and check the sleep influence distribution map 138 (c) for the date. The sleep quality indicator 138 (d) specifically shows the attribute value corresponding to the date to be selected from the user. Thus, the user may check specific attribute values of the sleep influencing factor according to the time quartile determined by the sleep evaluation model, for example, the degree of sleep influence. Meanwhile, the user may select a specific date from among the plurality of dates and check the sleep quality index 138 (d) for the date.

Hereinafter, the non-sleep state data used in the sleep analysis method and the sleep analysis device using the same will be described in detail with reference to FIGS. 6A and 6B.

6A illustrates non-sleep state data obtained from a wearable device, which is used in a sleep analysis method and a sleep analysis device using the same, according to an embodiment of the present invention.

Referring to FIG. 6A, non-sleep state data that may be measured from the wearable device may include data on activity amount, illuminance amount, HRV, and heart rate. More specifically, the non-sleep state data may include measurement date, user name, measurement time, X-axis, Y-axis, Z-axis, illuminance, X, Y, Z vector dot product representing the user's activity, HRV, and heart rate. Can be. In this case, the non-sleep state data measured from the wearable device may be input to the sleep quality prediction model. Thus, the sleep quality according to the non-sleep state data for the user may be predicted by the sleep quality prediction model.

Hereinafter, the sleep analysis method and sleep state data used in the sleep analysis device using the same according to an embodiment of the present invention will be described in detail with reference to FIG. 6B.

6B illustrates characteristic variables of non-sleep state data and sleep state data obtained from a wearable device, which are used in a sleep analysis method and a sleep analysis device using the same according to an embodiment of the present invention.

Referring to (a) of FIG. 6B, activity data during non-sleeping may be classified into various intensities and input to a sleep evaluation model. For example, the activity may be classified into upper, middle, and lower intensities based on the Metabolic Equivalent of Task (MET). More specifically, when the measured MET is less than 3, the activity amount is classified as 'low', when the measured MET is 3 or more and less than 6, the activity amount is 'medium', and when the measured MET is 6 or more, the activity amount is classified as 'high'. Can be. In this case, each of the first quartile, the second quartile, and the third quartile may mean a section from wake-up time to 12 pm, a section from 12 pm to 6 pm, and a section from 6 pm to bedtime. Thus, when the MET value according to each quartile during non-sleep is input to the sleep evaluation model for evaluating the sleep influencing factor, the characteristics of the quartiles affecting sleep may be output.

Referring to (b) of FIG. 6B, the illumination intensity data may be classified into the intensity of indoor low light, indoor high light and outdoor low light, and outdoor high light based on the measured LUX value and input to the sleep evaluation model. More specifically, when the measured illuminance is 0 or more and less than 320, the illuminance is 'indoor low illuminance', and when the measured illuminance is 320 or more and 500 or less, the illuminance is 'indoor high illuminance' and the measured illuminance is 501 or more and less than 10,000 The illuminance may be classified as 'outdoor low light', and when the measured illuminance is more than 10,000, the illuminance may be classified as 'outdoor high light'. In this case, each of the first quartile, the second quartile, and the third quartile may mean a section from wake-up time to 12 pm, a section from 12 pm to 6 pm, and a section from 6 pm to bedtime. Thus, when the intensity of illuminance according to each quartile at the time of non-sleeping is input to the sleep evaluation model for evaluating the sleep influencing factor, the characteristic affecting the sleep may be output.

Referring to FIG. 6B (c), the non-sleep state data and the sleep state data measured from the wearable device may include HRV data. The sleep analysis method and the device using the same according to an exemplary embodiment of the present invention may evaluate sleep disturbance factors by extracting features of non-sleep state data and sleep state data including HRV data.

More specifically, each of the HRV data measured in the non-sleep state and the HRV data measured in the sleep state may include a standard deviation of normal to normal heart rate (SDNN), a root mean square successive difference (RMSD) representing a degree of parasympathetic nerve, and sympathetic PNN50, which indicates the degree of nerve, and the extraction variable of mean heart rate. In the case of SDNN, the strength may be classified as 'good' if it is 0.06 or more and 'bad' if it is less than 0.06. In the case of RMSSD, if it is above 0.1, it may be classified as bad. Meanwhile, the sleep evaluation model may evaluate factors affecting sleep based on features extracted from HRV data.

Hereinafter, referring to FIG. 6C, a sleep analysis method and sleep recording data used in a sleep analysis device using the same according to an embodiment of the present invention will be described in detail.

6C illustrates sleep recording data used in a sleep analysis method and a sleep analysis device using the same according to an embodiment of the present invention.

Referring to FIG. 6C, the sleep recording data may be directly input by the user every day. At this time, the sleep record data, the recording date, name, sex, age, the time of the day, the time of sleep the day before, the quality of sleep the day before, the degree of freshness on the day, the wearing period of the wearable device, caffeine intake, alcohol intake, nap and BMI It may include.

Meanwhile, the non-sleep state data and the sleep state data used in various embodiments of the present invention may be obtained by various methods, without being limited to the above-described measurement method, and may be divided and used for high accuracy of prediction or evaluation. have. For example, non-sleep state data may include morning activity, afternoon activity, activity after sunset, indoor illuminance level, outdoor illuminance level, morning HRV, afternoon HRV, HRV after sunset, heart rate, and heart rate as measured by a user's non-sleep state. May include deviations. Furthermore, the state data during sleep may include HRV, heart rate, illuminance, sleep activity, and sleep efficiency measured in the user's sleep state.

In addition, the non-sleep state data and the sleep state data used in various embodiments of the present disclosure are not limited to the above-described features and feature values may be extracted in various ways. Furthermore, the sleep evaluation model may determine a sleep disturbance factor based on feature values extracted for the non-sleep state data and the sleep state data measured from the wearable device, but is not limited thereto. The sleep quality prediction model may predict sleep quality based on non-sleep state data measured from the wearable device, that is, raw raw data values, but is not limited thereto.

Example 1 Evaluation of a sleep evaluation model and a sleep quality prediction model used in various embodiments of the present invention

7A to 7C illustrate evaluation results of sleep disturbance factors based on a sleep evaluation model analyzed in a sleep analysis method and a sleep analysis system using the same according to an embodiment of the present invention.

Referring to FIG. 7A, the sleep evaluation model of the present invention may be a model configured to classify sleep influence factors affecting sleep from a plurality of decision trees, for example, a random forest, but is not limited thereto.

Referring back to FIG. 7A, the sleep evaluation model of the present invention for learning sleep influence factor determination may be a random forest composed of five decision trees. More specifically, in each decision tree in the learning phase, values of a learning feature vector, for example, learning non-sleep state data, may be input, and an operation of optimizing a parameter of a node partitioning function may be performed. At this time, the nodes forming the decision tree include a partition function, and each tree moves toward 0 (false) or 1 (true) lower node (child node) according to the result of the partition function, and the sleep influence factor is determined. Can be. In addition, each node may include a gini coefficient indicating the purity of the node and a value for a sample. According to the prediction result for each of the five decision trees, the sleep influencer decision tree 500 is finally determined.

Referring to FIG. 7B, when the values of the evaluation data are input to the sleep influence factor determination tree 500, the sleep influence factor is determined by the division function of each node described above (orange box direction), and the determined sleep influence factor is determined. The results for (good or bad), and the accuracy rate are provided together.

Referring to (a) of FIG. 7C, 462 predetermined trading sets were used to train the sleep evaluation model, and the same 462 data was used as the evaluation data. At this time, the sleep diary of the sleep recording data was not considered in the evaluation. As a result, according to the sleep evaluation model of the present invention, the accuracy of prediction was 83.55%, the sensitivity was 99.46%, and the specificity was 16.85%. That is, the sleep evaluation model of the present invention can evaluate factors influencing sleep with high accuracy and sensitivity, and thus can be applied to a sleep analysis system.

Referring to (b) of FIG. 7C, 462 predetermined trading sets were used to train the sleep evaluation model, and the same 462 data was used as the evaluation data. In this case, the sleep diary was used as sleep record data for evaluating sleep influencing factors in this evaluation. As a result, according to the sleep evaluation model of the present invention, the accuracy of prediction was 83.12%, the sensitivity was 100%, and the specificity was 12.36%.

That is, the sleep evaluation model of the present invention can evaluate factors influencing sleep with high accuracy and sensitivity, and thus can be applied to a sleep analysis system. Furthermore, the sleep evaluation model of the present invention can evaluate factors that affect sleep at an excellent level when using sleep record data such as a sleep diary created by a user.

7D illustrates a sleep quality prediction result based on a sleep quality prediction model analyzed in a sleep analysis method and a sleep analysis system using the same according to an embodiment of the present invention.

In this case, the sleep quality prediction model of the present invention may be a model based on a deep learning algorithm of DNN, CNN, DCNN, RNN, RBM, DBN or SSD, but is not limited thereto.

Referring to (a) of FIG. 7D, the accuracy of the sleep quality prediction for the training data set including the non-sleep state data for the sleep quality prediction model of the present invention is about 93%. For the sleep quality prediction model of the present invention, the accuracy of sleep quality prediction for the evaluation data set consisting of non-sleep state data is about 87%.

More specifically, referring to (b) of FIG. 7D, the convolutional layer 4 of the sleep quality prediction model of the present invention configured to predict sleep quality is a value of a batch representing a sample number before parameter updating. The size was set to 5, the number of epochs representing the number of trainings was set to 10, the number of filters was set to 50, and the kernel_size was set to 2. As a result, the sleep quality prediction accuracy for the learning data set was 93.37%, and the sleep quality prediction accuracy for the evaluation data was 87.23%.

According to the results of the first embodiment, the sleep evaluation model and the sleep quality prediction model used in various embodiments of the present invention can predict the sleep influencing factor and the sleep quality with high accuracy and sensitivity. In particular, the above results may mean that data on activity during sleep, illuminance, heart rate and HRV are deeply related to sleep.

Accordingly, the present invention can evaluate and provide factors that affect the sleep of the user based on non-sleep state data measured from the wearable device, more specifically, activity, illuminance, heart rate, and HRV data. Quality can be predicted and provided.

Furthermore, the present invention has the effect of solving problems such as expensive analysis cost and inconvenience that a user has in the conventional sleep analysis system, and recognizes the cause of insomnia for a user with a sleep disorder, and improves it. It has the effect of providing information to help.

Each of the features of the various embodiments of the present invention may be combined or combined with each other in part or in whole, various technically interlocking and driving as can be understood by those skilled in the art, each of the embodiments may be implemented independently of each other It may be possible to carry out together in an association.

Although the embodiments of the present invention have been described in more detail with reference to the accompanying drawings, the present invention is not necessarily limited to these embodiments, and various modifications can be made without departing from the spirit of the present invention. . Therefore, the embodiments disclosed in the present invention are not intended to limit the technical idea of the present invention but to describe the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments. Therefore, it should be understood that the embodiments described above are exemplary in all respects and not restrictive. The protection scope of the present invention should be interpreted by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present invention.

100: sleep analysis device
110: receiver
120: input unit
130: display unit
132: menu
134: What is data management
136: Time series sleep analysis column
136 (a): Time series state analysis diagram
136 (b): Section selection bar
138: Sleep distribution chart
138 (a): Distribution map of sleep quality prediction
138 (b): sleep impact
138 (c): Surface distribution map
140: storage unit
150: processor
200: wearable device
300: bed fixed device
400: medical staff device
500: Sleep Influencer Decision Tree
1000: sleep analysis system

Claims (24)

A sleep analysis method performed by a sleep analysis device comprising a receiver and a processor,
Receiving, from the wearable device, non-sleep state data and sleep state data for the user through the receiving unit;
Evaluating, by the processor, an influence factor of sleep on the user using a sleep evaluation model configured to evaluate an influence factor of sleep based on the non-sleep state data and the sleep state data;
The receiving step,
Receiving each of the non-sleep state data and the sleep state data through a plurality of time quartiles through the receiving unit,
Evaluating the influence factor of the sleep,
Evaluating, by the processor, the influencing factor of sleep according to time quartile based on the non-sleep state data and the sleep state data received in the plurality of time quartiles using the sleep evaluation model. , Sleep analysis method. delete The method of claim 1,
Evaluating the influence factor of the sleep,
And calculating, by the processor, a relative or absolute degree of sleep influence on factors influencing sleep according to the time quartile. The method of claim 3,
Performed after the step of evaluating the influencing factor of sleep,
Generating, by the processor, a sleep impact degree configured to have an area size proportional to the sleep influence degree corresponding to the influence factor of sleep according to the time quartile; and
And providing the sleep impact. The method of claim 3,
Performed after the step of evaluating the influencing factor of sleep,
Generating, by the processor, a sleep influence distribution map configured to have a distribution degree proportional to the sleep influence degree corresponding to the influence factor of sleep according to the time quartile; and
Providing the sleep impact distribution map. The method of claim 1,
The receiving step,
Receiving, through the receiving unit, non-sleep state data and the sleep state data for each of the plurality of days at the plurality of time quartiles,
Evaluating factors affecting sleep according to the time quartile,
Influence of sleep over time by the processor, based on the non-sleep state data and the sleep state data received in the plurality of time quartiles for each of the plurality of dates, using the sleep assessment model. Assessing the factors,
After the step of evaluating the factors influencing sleep according to the time quartile,
Receiving a specific date selected from the plurality of dates through the receiving unit, and
And providing an influencing factor of sleep according to the time quartile corresponding to the specific date. The method of claim 1,
The receiving step,
Receiving, from the wearable device, the non-sleep state data and the sleep state data for a predetermined period through the receiving unit;
Performed after the receiving step,
Generating, by the processor, a time series state analysis diagram for each of the non-sleep state data and the sleep state data for the predetermined period of time, and
Providing the time series state analysis diagram. The method of claim 1,
Performed after the receiving step,
Extracting, by the processor, a feature variable for each of the non-sleep state data and the sleep state data,
Evaluating the influence factor of the sleep,
Evaluating, by the processor, an influencing factor of sleep on the user using the sleep assessment model based on the feature variable for each of the non-sleep state data and the sleep state data. Analytical Method. The method of claim 1,
Performed after the receiving step,
Predicting, by the processor, sleep quality of the user using a sleep quality prediction model configured to predict sleep quality based on the non-sleep state data; and
And providing a plurality of levels of the predicted quality of sleep. The method of claim 1,
Evaluating the influence factor of the sleep,
And calculating, by the processor, a sleep efficiency of the user based on the sleep state data using the sleep assessment model. The method of claim 1,
The non-sleep state data,
And at least one of a group consisting of activity amount, illuminance amount, heart rate, and HRV measured by the wearable device in a non-sleep state of the user. The method of claim 1,
The sleep state data,
Sleep measurement data measured by the wearable device in the sleep state of the user, and sleep recording data input every predetermined period from the user. The method of claim 12,
The measurement data when sleeping,
Sleep at least one of the group consisting of HRV, heart rate, illuminance and sleep activity measured in the sleep state of the user. The method of claim 12,
The sleep recording data,
At least one selected from a date, a name, a gender, an age, a day that happened on the day, a day of sleep, a day's sleep quality, a day's freshness, an unwearing period of the wearable device, a caffeine intake, an alcohol intake, a nap, and a BMI. One, sleep analysis method. A receiver configured to receive non-sleep state data and sleep state data for the user from the wearable device, and
A processor connected to the receiver,
The processor,
And use the sleep evaluation model configured to evaluate the influence factor of sleep based on the non-sleep state data and the sleep state data, to evaluate the influence factor of sleep on the user,
The receiving unit,
Receive non-sleep state data and each of the sleep state data in a plurality of time quartiles,
The processor,
And use the sleep evaluation model to evaluate factors affecting sleep over time quartile based on the non-sleep state data and the sleep state data received in the plurality of time quartiles. delete The method of claim 15,
The processor,
And calculate a degree of relative or absolute sleep influence on the influencing factor of sleep over the time quartile. The method of claim 17,
The processor,
Generate a sleep influence map configured to have an area size proportional to the sleep influence degree corresponding to the influence factor of sleep according to the time quartile,
The sleep analysis device, further configured to provide the sleep impact. The method of claim 17,
The processor,
And generate a sleep influence distribution plot configured to have a distribution degree proportional to the sleep influence degree corresponding to the influence factor of sleep over the time quartile and to provide the sleep influence distribution map. The method of claim 15,
The receiving unit,
For each of a plurality of dates, the non-sleep state data and the sleep state data in the plurality of time quartiles, and
Is further configured to receive a selected specific date of the plurality of dates,
The processor,
Using the sleep evaluation model, based on the non-sleep state data and the sleep state data received in the plurality of time quartiles for each of the plurality of dates, the influence factor of sleep over time quartiles is evaluated, The sleep analysis device is further configured to provide an influencing factor of sleep according to the time quartile corresponding to the specific date. The method of claim 15,
The receiving unit,
Is further configured to receive the non-sleep state data and sleep state data for a predetermined period from a wearable device,
The processor,
And generate a time series state analysis plot for each of the non-sleep state data and sleep state data for the predetermined period of time, and provide the time series state analysis plot. The method of claim 15,
The processor,
Further configured to predict sleep quality of the user by using a sleep quality prediction model configured to predict sleep quality based on the non-sleep state data, and to provide the predicted sleep quality at a plurality of levels, Sleep analysis device. The method of claim 15,
The processor,
And use the sleep assessment model to calculate a sleep efficiency of the user based on the sleep state data. The method of claim 15,
The processor,
And extract feature variables for each of the non-sleep state data and the sleep state data,
The sleep evaluation model,
And further evaluate an influencing factor of sleep on the user based on the non-sleep state data and the feature variable for each of the sleep state data.

Images