[go: up one dir, main page]
More Web Proxy on the site http://driver.im/

KR20140016559A - Apparatus and method for monitoring disease status based on time-series data modeling - Google Patents

Apparatus and method for monitoring disease status based on time-series data modeling Download PDF

Info

Publication number
KR20140016559A
KR20140016559A KR1020120083271A KR20120083271A KR20140016559A KR 20140016559 A KR20140016559 A KR 20140016559A KR 1020120083271 A KR1020120083271 A KR 1020120083271A KR 20120083271 A KR20120083271 A KR 20120083271A KR 20140016559 A KR20140016559 A KR 20140016559A
Authority
KR
South Korea
Prior art keywords
model
data
disease
time series
analysis
Prior art date
Application number
KR1020120083271A
Other languages
Korean (ko)
Inventor
감혜진
김하영
Original Assignee
삼성전자주식회사
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 삼성전자주식회사 filed Critical 삼성전자주식회사
Priority to KR1020120083271A priority Critical patent/KR20140016559A/en
Publication of KR20140016559A publication Critical patent/KR20140016559A/en

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders

Landscapes

  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Pathology (AREA)
  • Primary Health Care (AREA)
  • Epidemiology (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Biophysics (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • Veterinary Medicine (AREA)
  • Measuring And Recording Apparatus For Diagnosis (AREA)

Abstract

A device for monitoring disease status by analyzing structural changes in time series data. According to an embodiment, the apparatus for monitoring a disease state may estimate a parameter of an analysis model for each structural change section with respect to a data processing unit for obtaining time series data from sensor data, and at least one structural change section partitioned by structural change points in the time series data. And a model constructing unit for constructing a classification model for determining a change in a disease state based on the data analyzing unit and its parameters.

Figure P1020120083271

Description

Apparatus and method for monitoring disease status based on time series data modeling {APPARATUS AND METHOD FOR MONITORING DISEASE STATUS BASED ON TIME-SERIES DATA MODELING}

A device and method for monitoring disease status by analyzing structural changes in time series data.

For the majority of chronic diseases, monitoring the status changes of the patient's disease invention, progress, and improvement is very important in the treatment, management and prevention of the disease. In addition to changes in the patient's condition due to the disease itself, the effects of intentional interventions such as drug administration, exercise regimen, surgery, and unintentional interventions such as changes in the patient's physical or psychological state visit the hospital. In addition, it cannot be easily identified by a short round of hospitalization, etc., and it is necessary to continuously monitor and analyze the pattern of variation of long time series observations.

Analysis of patient conditions using Actigraphs has been conducted since the 1980s for a variety of conditions, including ADHD, dementia, sleep disorders, depression, cancer, heart disease, and drug addiction. However, previous studies have extracted some characteristic values from the actigraph and suggested statistical differences between the disease / non-disease groups based on activity change data in actual patients. It is not enough to model changes in conditions such as improvement or to predict future disease states.

An apparatus and method are provided for analyzing changes in activity structure through modeling of time series activity data to provide time and characteristic information of disease state changes.

According to an aspect, the apparatus for monitoring a disease state may include a data processor obtaining time series data from sensor data, and estimating a parameter of an analysis model for each structural change section with respect to one or more structural change sections partitioned by structural change points in the time series data. It may include a model construction unit for constructing a classification model for determining a change in the state of the disease based on the data analysis unit and the parameters.

In this case, the sensor data may be continuous measurement data measured for a predetermined period by one or more sensors including an acceleration sensor.

The data processor may determine the analysis section according to the characteristics of the disease from the sensor data, and obtain time series data by extracting the sensor data of the analysis section.

The data processor may determine the analysis section based on the characteristic of the disease in the sensor data, and calculate time-series data by calculating a feature value based on the sensor data of the analysis section.

In this case, the feature value may include one or more of the sum of data, the mean, the median value, the maximum value, the minimum value, the variance, the standard deviation, the number of outliers, the number of data above the reference value, the number of data below the reference value, and the derivative value.

The data analysis unit may select an analysis model for each structural change section from the time series data analysis model through a model selection method including a Bayesian Information Criterion (BIC).

The time series data analysis model includes a Time Varying Coefficient (TVC) model, an AutoRegressive Conditional Heteroskedasticity (ARCH) model, a Generalizes ARCH (GARCH) model, a Stochastic Volatility (SV) model, an AutoRegressive Integrated Moving Average (ARIMA) model, and one or more of the models. It can include a combined model.

The data analyzer may estimate the parameters of the analysis model using any one of estimation techniques including statistical hypothesis testing, least squares principle, and dynamic algorithm.

The data analyzer may detect the structural change point through a detection technique including statistical hypothesis testing and RSS (residual squared sum) using any one model selected from a time series data analysis model.

According to an additional aspect, the disease state monitoring apparatus may further include a monitoring unit that determines a change in the state of the disease based on the established classification model when the monitoring data about the disease is input.

According to an aspect, the disease state monitoring method may include obtaining time series data from sensor data, and estimating a parameter of an analysis model for each structural change section with respect to one or more structural change sections partitioned by structural change points in the time series data. And building a classification model for determining a change in state of the disease based on the parameter.

In this case, the sensor data may be continuous measurement data measured for a predetermined period by one or more sensors including an acceleration sensor.

In the time series data obtaining step, the analysis section may be determined according to the characteristics of the disease in the sensor data, and the time series data may be obtained by extracting the sensor data of the analysis section.

In addition, the time series data acquisition step may determine the analysis section according to the characteristics of the disease in the sensor data and calculate the feature value based on the sensor data of the analysis section to obtain time series data.

In this case, the feature value may include one or more of the sum of data, the mean, the median value, the maximum value, the minimum value, the variance, the standard deviation, the number of outliers, the number of data above the reference value, the number of data below the reference value, and the derivative value.

According to an additional aspect, the disease state monitoring method may further include selecting an analysis model for each structural change section from a time series data analysis model through a model selection method including a Bayesian Information Criterion (BIC).

The time series data analysis model includes a Time Varying Coefficient (TVC) model, an AutoRegressive Conditional Heteroskedasticity (ARCH) model, a Generalizes ARCH (GARCH) model, a Stochastic Volatility (SV) model, an AutoRegressive Integrated Moving Average (ARIMA) model, and one or more of the models. It may include a combined model.

The parameter estimating step may estimate the parameters of the analysis model using any one of estimation techniques including statistical hypothesis testing, Least Squares Principle, and dynamic algorithm.

According to a further aspect, the disease state monitoring method may further include detecting a structural change point through a detection technique including a statistical hypothesis test and a residual squared sum (RSS) using any one model selected from a time series data analysis model. It may include.

According to an additional aspect, the disease state monitoring method may further include determining a state change of the disease based on the established classification model when monitoring data about the disease is input.

By analyzing changes in activity structure through modeling of time series data, the disease state can be monitored and the timing and characteristics of the disease state change can be provided.

1 is a block diagram of a disease state monitoring apparatus according to an embodiment.
2A to 2C are examples for describing a disease state monitoring apparatus according to the embodiment of FIG. 1.
3 is a block diagram of a disease state monitoring apparatus according to another embodiment.
4 is a flowchart of a disease state monitoring method according to an embodiment.
5 is a flowchart of a disease state monitoring method according to another embodiment.

The details of other embodiments are included in the detailed description and drawings. BRIEF DESCRIPTION OF THE DRAWINGS The advantages and features of the present invention, and the manner of achieving them, will be apparent from and elucidated with reference to the embodiments described hereinafter in conjunction with the accompanying drawings. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. To fully disclose the scope of the invention to those skilled in the art, and the invention is only defined by the scope of the claims. Like reference numerals refer to like elements throughout the specification.

Hereinafter, an apparatus and method for monitoring disease state based on time series data modeling according to embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram of a disease state monitoring apparatus according to an embodiment.

Referring to FIG. 1, the disease state monitoring apparatus 100 may include a data processor 110, a data analyzer 120, and a model builder 130.

The data processing unit 110 uses a variety of sensors such as an acceleration sensor for measuring activity data, a biosensor for measuring blood sugar, blood pressure, and a stress test sensor for measuring a stress level (eg, 1 day, 1 week, 1). Monthly, one year, etc.) can be collected continuously measured sensor data. In this case, the data collection period may be set differently according to the characteristics of the disease. For example, for a rapidly changing state of ADHD or Parkinson's disease, the collection period can be several seconds or minutes, and for depression that shows a mild change and months of recurrence, months can be set as the collection period. Can be.

The data processor 110 may acquire time series data for analyzing a disease state using the collected sensor data. The data processor 110 may determine an analysis section based on the characteristics of the disease and use the sensor data of the analysis section as time series data as it is. The analysis section may be the entire collection section of the sensor data, or may be a clinically meaningful section of the sensor data (eg, night, day, sleep, 3 hours after taking the medicine, etc.).

The data processor 110 may process the sensor data of the analysis section to be suitable for the analysis of the disease state, extract a feature value, and obtain time series data using the feature value. In this case, the feature value is the sum, mean, median, maximum, minimum, variance, and standard deviation of the data calculated for each predetermined time unit (eg, seconds, minutes, hours, days, weeks, etc.) using the sensor data of the analysis section. , The number of outliers, the number of data above the reference value, the number of data below the reference value, and the derivative value.

The data analyzer 120 may detect a structural change point in the obtained time series data, and select an analysis model for each section with respect to the structural change section partitioned by the structural change point. The structural change point may be detected using a detection hypothesis including statistical hypothesis testing and a residual squared sum (RSS) using any one model selected from a time series data analysis model. In addition, when the structural change point can be clearly identified with the naked eye, information about the structural change point may be input from the user.

As the analysis model for each section, the model selected for detecting the structural change point can be used as it is. In this case, the analysis time can be saved by detecting the structural change point together with the analysis model selection. Alternatively, when a structural change point is detected, an appropriate analysis model may be selected from a time series data analysis model for each structural change section. The selected analysis model may be the same or different for each structural change section.

The data analyzer 120 may select a model or an analysis model for detecting a structural change point from the time series data analysis model through a model selection method including a Bayesian Information Criterion (BIC). Here, the time series data analysis model includes a Time Varying Coefficient (TVC) model, an AutoRegressive Conditional Heteroskedasticity (ARCH) model, a Generalizes ARCH (GARCH) model, a Stochastic Volatility (SV) model, an AutoRegressive Integrated Moving Average (ARIMA) model, and one of these models. It can include a combination of the above models.

The data analyzer 120 may estimate a parameter of an analysis model for each structural change section selected for at least one structural change section in the time series data. The data analyzer 120 may estimate a parameter of the analysis model using any one of estimation techniques including a statistical hypothesis test, a Least Squares Principle, and a dynamic algorithm.

For example, Equation 1 below shows a general formula without structural change when the GARCH (p, q) model is called a detection model or an analysis model.

Figure pat00001

here,

Figure pat00002
Is independent and is an arbitrary variable with a normal distribution with mean 0 and variance 1.

If there is a structural change of m + 1 in the analytical model and the time points are T 0 , T 1 , ... T m +1 , respectively, the parameters (α 0 , α t , β k ) are different at each point. Therefore, Equation 1 may be expressed as Equation 2 below.

Figure pat00003

That is, if the GARCH (p, q) model is a structural change point detection model, the data analysis unit 120 indicates the number of points where the structural change occurs, m + 1 and the time points T 0 , T 1 , ... T m. +1 can be detected. Also, If the GARCH (p, q) model is an analytic model for a structural change interval, the parameters (α 0 , α t , β k ) of each analytical model are tested by statistical hypothesis, least squares principle, and dynamic algorithm. algorithm).

The model building unit 130 may build a classification model for determining a change in the state of the disease by using the estimated parameter. Meanwhile, when the feature value is extracted by the data processor 110, the feature value may be used in constructing a classification model together with the estimated parameter. In this case, the classification model may be a classification rule that associates parameters with characteristics of the disease, such as the stage of the disease, whether the disease is developed or cured, or worsens or alleviates the disease. That is, when the parameter increases or decreases according to the severity of the disease, when the parameter of the newly input monitoring data decreases, a rule may be determined that the disease state of the patient improves. Meanwhile, the classification model may be constructed for each individual or one classification model may be constructed by deriving common features for a plurality of patients for each disease.

According to an additional aspect, the disease state analysis apparatus 100 may further include a model DB 150. The classification model built by the model building unit 130 may be stored in the model DB 150 to be utilized in future disease state monitoring.

When new monitoring data is input, the monitoring unit 140 may determine a change in the state of the disease by using the established classification model. In this case, the determination result may be provided to the user.

2A to 2C are examples for describing a disease state monitoring apparatus according to the embodiment of FIG. 1.

2A illustrates the obtained time series data. The data processor 110 may acquire time series data as shown in FIG. 2 using the collected sensor data. The acquired time series data shows a similar pattern repeatedly or in a form that is not visible to the naked eye.

FIG. 2B illustrates three structural change intervals 10a, 10b, and 10c in which the structural change points are derived and partitioned by the data analyzer 120 from the time series data illustrated in FIG. 2. The numbers assigned to each of the structural change sections 10a, 10b, and 10c in FIG. 2b illustrate the stages of the disease in each of the structural change sections 10a, 10b, and 10c. For example, when the disease to be analyzed is depression, the numbers 3, 2, and 1 may mean that the depression is a very serious state, a little severe state, and a little out of normal state, respectively. That is, the structural change section 10a of FIG. 2b is a case where the patient is in a very severe depression state, and the structural change section 10b is a case where the depression state is improved from a very severe state to a slightly severe state, and the structural change section 10c is a case of the patient. Depression can mean an almost normal level.

When the model builder 130 estimates the parameters by the data analyzer 120, the parameters of the structural change sections 10a, 10b, and 10c and the stages of the disease (3, 2, 1) are illustrated in FIG. 2B. It is possible to build a classification model that can match. For example, the patient's depression level can be measured using clinically widely used depression measurement scales such as the Beck Depression Inventory (BDI), Hamilton rating scale for depression (HAM-D), or Hamilton anxiety rating scale (HAMA). Based on the diagnosis and the like, it is possible to establish a matching relationship between parameters and depression stages for each structural change section.

2C, when new data for monitoring a disease is input, the monitoring unit 130 compares the parameters of the analysis model with respect to the structural change intervals 20a and 20b of the monitoring data and the constructed classification model to compare each structural change interval. The disease stage of (20a, 20b) can be determined, and the change in the disease state can be determined. That is, referring to FIG. 2C, the monitoring unit 130 may determine the disease stage as 2 and 3, respectively, through the classification model constructed for each structural change section 20a and 20b. , 2, 1 to 2, 3, it can be determined that the disease state improves and then worsens again. The monitoring unit 130 may provide the user with appropriate information based on the determination result of the disease state change.

According to the disclosed embodiments, it is possible to analyze and provide the characteristics of the timing and condition of the physical or psychological state change of the patient based on the structural change information, thereby preventing or early detecting the recurrence of the disease of high risk of recurrence can do.

3 is a block diagram of a disease state monitoring apparatus according to another embodiment.

Referring to FIG. 3, the disease state monitoring apparatus 200 according to another embodiment includes a data processor 210, a data analyzer 220, a monitor 230, and a model DB 240.

The data processor 210 may obtain time series data through the monitoring data when the monitoring data about the disease is input. In this case, the monitoring data may be sensor data collected through various sensors such as an acceleration sensor and a biosensor. The data processor 210 may acquire time series data through the monitoring data as described above.

The data analyzer 220 may select a structural change point of the time series data and an analysis model for each structural change section. In addition, parameters for the analysis model for each structural change section are estimated.

The monitoring unit 230 may search for a corresponding classification model that is already built in the model DB 240, and compare the classification model with the estimated parameter to monitor the change in the state of the disease. That is, the monitoring unit 230 matches the parameters estimated for each structural change interval with each disease stage through the classification model searched as illustrated in FIG. 2C, and analyzes the change in disease state by analyzing the change in disease stage. You can judge.

4 is a flowchart of a disease state monitoring method according to an embodiment.

Referring to FIG. 4, first, the disease state monitoring apparatus 100 obtains time series data for analyzing a disease state using sensor data (step 301). The disease state monitoring apparatus 100 may collect sensor data continuously measured for a predetermined period (eg, 1 day, 1 week, 1 month, 1 year, etc.) through various sensors such as an acceleration sensor and a bio sensor.

The disease state monitoring apparatus 100 may determine an analysis section from time series data according to a characteristic of a disease, and use the sensor data of the analysis section or a feature value extracted to be suitable for disease state analysis as time series data. In this case, the analysis section may be the entire collection section of the sensor data, or may be a clinically meaningful section of the sensor data (eg, night, day, during sleep, 3 hours after taking the medicine, etc.). have. In addition, the feature value is the total, average, median, maximum, minimum, variance, standard deviation, and outlier for each predetermined time unit (e.g., seconds, minutes, hours, days, weeks, etc.) using the sensor data of the analysis section. Number, data above the reference value, data below the reference value, and derivative value.

The disease state monitoring apparatus 100 may then detect the point of structural change in the obtained time series data (step 302). The disease state monitoring apparatus 100 may detect a structural change point through a detection technique including a statistical hypothesis test and a residual squared sum (RSS) using any one model selected from a time series data analysis model. In this case, the time series data analysis model includes a Time Varying Coefficient (TVC) model, an AutoRegressive Conditional Heteroskedasticity (ARCH) model, a Generalizes ARCH (GARCH) model, a Stochastic Volatility (SV) model, an AutoRegressive Integrated Moving Average (ARIMA) model, and one of these models. The model may be a combination of the above, and the model selected for detecting the structural change point may be a regression model or an autoregressive integrated moving average (ARIMA).

Next, an analysis model for each section may be selected for each structural change section partitioned by the structural change point (step 303). The analysis model for each section may use the model selected for detecting structural change points. Alternatively, when a structural change point is detected, an appropriate analysis model may be selected from a time series data analysis model for each structural change section.

Then, the parameters of the selected analysis model are estimated for each structural change section (step 304). In this case, the parameter may be estimated using any one of estimation techniques including a statistical hypothesis test, a Least Squares Principle, and a dynamic algorithm.

Then, using the estimated parameters, a classification model may be constructed to determine a change in the state of the disease (step 305). If there are feature values extracted when time series data is obtained, the feature values may be used in constructing a classification model together with the estimated parameters.

Finally, when new monitoring data is input, the disease state monitoring apparatus 100 may determine a change in the state of the disease by using the established classification model (step 306).

5 is a flowchart of a disease state monitoring method according to another embodiment.

Referring to FIG. 5, when monitoring data for a disease is input, the disease state monitoring apparatus 200 may obtain time series data through the monitoring data (step 401).

Next, the disease state monitoring apparatus 200 may detect a structural change point of the time series data (step 402), and select an analysis model for each structural change section (step 403).

Next, parameters for the analysis model for each structural change section are estimated (step 404).

Finally, the model DB 240 may search for a corresponding classification model that is already built, compare the classification model with the estimated parameters, and monitor the change in the state of the disease (step 405).

Meanwhile, the embodiments of the present invention can be embodied as computer readable codes on a computer readable recording medium. A computer-readable recording medium includes all kinds of recording apparatuses in which data that can be read by a computer system is stored.

Examples of the computer-readable recording medium include a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device and the like, and also a carrier wave (for example, transmission via the Internet) . The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion. And functional programs, codes and code segments for implementing the present invention can be easily inferred by programmers in the art to which the present invention belongs.

It will be understood by those skilled in the art that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive. The scope of the present invention is defined by the appended claims rather than the foregoing detailed description, and all changes or modifications derived from the meaning and scope of the claims and the equivalents thereof are included in the scope of the present invention Should be interpreted.

100: disease state monitoring device 110: data processing unit
120: data analysis unit 130: model construction unit
140: monitoring unit 150: model DB

Claims (20)

A data processor for obtaining time series data from sensor data;
A data analyzer for estimating a parameter of an analysis model for each structural change section with respect to at least one structural change section partitioned by structural change points in the time series data; And
And a model building unit for constructing a classification model for determining a change in a state of a disease based on the parameter. 2. The method according to claim 1,
A device for monitoring a disease state which is continuous measurement data measured for a predetermined period of time by at least one sensor including an acceleration sensor. The method of claim 2, wherein the data processing unit,
A disease state monitoring apparatus for determining an analysis section according to a characteristic of a disease from the sensor data, and extracting sensor data of the analysis section to obtain time series data. The method of claim 2, wherein the data processing unit,
The apparatus for monitoring disease states according to the sensor data, determining an analysis section according to a characteristic of a disease, and obtaining time series data by calculating a feature value based on sensor data of the analysis section. The method of claim 4, wherein the feature value
A device for monitoring a disease state including at least one of a sum, mean, median value, maximum value, minimum value, variance, standard deviation, outlier number, data above a reference value, data below a reference value, and derivative values. The method of claim 1, wherein the data analysis unit,
A disease state monitoring apparatus for selecting an analysis model for each structural change section from a time series data analysis model through a model selection method including a Bayesian Information Criterion (BIC). The method of claim 6, wherein the time series data analysis model,
Including a Time Varying Coefficient (TVC) model, an AutoRegressive Conditional Heteroskedasticity (ARCH) model, a Generalizes ARCH (GARCH) model, a Stochastic Volatility (SV) model, an AutoRegressive Integrated Moving Average (ARIMA) model, and a combination of one or more of the above models Disease condition monitoring device. The method of claim 1, wherein the data analysis unit,
A disease state monitoring apparatus for estimating the parameters of the analysis model using any one of estimation techniques including statistical hypothesis testing, Least Squares Principle, and dynamic algorithm. The method of claim 1, wherein the data analysis unit,
A disease state monitoring apparatus for detecting the structural change point through a detection technique including statistical hypothesis testing and RSS (residual squared sum) using any one model selected from a time series data analysis model. The method of claim 1,
And a monitoring unit configured to determine a change in the state of the disease based on the established classification model when the monitoring data about the disease is input. Obtaining time series data from sensor data;
Estimating a parameter of an analysis model for each structural change section for at least one structural change section partitioned by structural change points in the time series data; And
Constructing a classification model for determining a change in the condition of the disease based on the parameter. The method of claim 11, wherein the sensor data,
A method for monitoring disease state that is continuous measurement data measured for a predetermined period of time by one or more sensors including an acceleration sensor. The method of claim 12, wherein the time series data obtaining step comprises:
The disease state monitoring method of determining the analysis section according to the characteristics of the disease in the sensor data and to obtain time series data by extracting the sensor data of the analysis section. The method of claim 12, wherein the time series data obtaining step comprises:
A disease state monitoring method of determining an analysis section according to a characteristic of a disease from the sensor data and obtaining time series data by calculating a feature value based on sensor data of the analysis section. The method of claim 14, wherein the feature value is
A disease state monitoring method comprising at least one of a sum of data, an average, a median, a maximum value, a minimum value, a variance, a standard deviation, an outlier number, a data above a reference value, a data below a reference value, and a derivative. 12. The method of claim 11,
Selecting an analysis model for each structural change section from a time series data analysis model through a model selection method including a Bayesian Information Criterion (BIC). The method of claim 16, wherein the time series data analysis model,
Including a Time Varying Coefficient (TVC) model, an AutoRegressive Conditional Heteroskedasticity (ARCH) model, a Generalizes ARCH (GARCH) model, a Stochastic Volatility (SV) model, an AutoRegressive Integrated Moving Average (ARIMA) model, and a combination of one or more of the above models How to monitor disease status. The method of claim 11, wherein the parameter estimating step comprises:
A disease state monitoring method for estimating the parameters of the analytical model using any one of estimation techniques including statistical hypothesis testing, least squares principle, and dynamic algorithm. 12. The method of claim 11,
Detecting the structural change point through a detection technique including statistical hypothesis testing and RSS (residual squared sum) using any one model selected from a time-series data analysis model. 12. The method of claim 11,
If the monitoring data for the disease is input, determining a change in the condition of the disease based on the established classification model; disease state monitoring method further comprising.

KR1020120083271A 2012-07-30 2012-07-30 Apparatus and method for monitoring disease status based on time-series data modeling KR20140016559A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
KR1020120083271A KR20140016559A (en) 2012-07-30 2012-07-30 Apparatus and method for monitoring disease status based on time-series data modeling

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1020120083271A KR20140016559A (en) 2012-07-30 2012-07-30 Apparatus and method for monitoring disease status based on time-series data modeling

Publications (1)

Publication Number Publication Date
KR20140016559A true KR20140016559A (en) 2014-02-10

Family

ID=50265520

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1020120083271A KR20140016559A (en) 2012-07-30 2012-07-30 Apparatus and method for monitoring disease status based on time-series data modeling

Country Status (1)

Country Link
KR (1) KR20140016559A (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101956717B1 (en) * 2017-09-01 2019-03-11 군산대학교산학협력단 Method and apparatus for prediction of wind direction, and method for yaw control of wind turbines using the same
KR101956715B1 (en) * 2017-09-01 2019-03-11 군산대학교산학협력단 Wind direction prediction method and apparatus for yaw control of wind turbines
CN114383646A (en) * 2021-10-29 2022-04-22 廊坊市大华夏神农信息技术有限公司 Method and equipment for detecting resolution of continuously-changing type measured sensor
US11636377B1 (en) * 2018-07-03 2023-04-25 Amazon Technologies, Inc. Artificial intelligence system incorporating automatic model updates based on change point detection using time series decomposing and clustering
US11651271B1 (en) 2018-07-03 2023-05-16 Amazon Technologies, Inc. Artificial intelligence system incorporating automatic model updates based on change point detection using likelihood ratios

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101956717B1 (en) * 2017-09-01 2019-03-11 군산대학교산학협력단 Method and apparatus for prediction of wind direction, and method for yaw control of wind turbines using the same
KR101956715B1 (en) * 2017-09-01 2019-03-11 군산대학교산학협력단 Wind direction prediction method and apparatus for yaw control of wind turbines
US11636377B1 (en) * 2018-07-03 2023-04-25 Amazon Technologies, Inc. Artificial intelligence system incorporating automatic model updates based on change point detection using time series decomposing and clustering
US11651271B1 (en) 2018-07-03 2023-05-16 Amazon Technologies, Inc. Artificial intelligence system incorporating automatic model updates based on change point detection using likelihood ratios
CN114383646A (en) * 2021-10-29 2022-04-22 廊坊市大华夏神农信息技术有限公司 Method and equipment for detecting resolution of continuously-changing type measured sensor
CN114383646B (en) * 2021-10-29 2023-08-25 廊坊市大华夏神农信息技术有限公司 Method and equipment for detecting resolution of continuously-variable measured sensor

Similar Documents

Publication Publication Date Title
EP3468450B1 (en) Method and system for analyzing human gait
US8823526B2 (en) Method of assessing human fall risk using mobile systems
Quante et al. Practical considerations in using accelerometers to assess physical activity, sedentary behavior, and sleep
US11357410B2 (en) Measuring blood pressure
O’Connor et al. Automatic detection of gait events using kinematic data
US11039760B2 (en) Detection of walking in measurements of the movement of a user
US20120221310A1 (en) System for analyzing physiological signals to predict medical conditions
KR20170023770A (en) Diagnosis model generation system and method
US9204797B2 (en) Gait-based biometric system for detecting pathomechanical abnormalities relating to disease pathology
Thorpe et al. Development of a sensor-based behavioral monitoring solution to support dementia care
CN108027363A (en) It is engineered haemocyte estimation
US10885041B2 (en) Gait-based biometric data analysis system
JP6659049B2 (en) Blood sugar level prediction device, blood sugar level prediction method and program
KR20140016559A (en) Apparatus and method for monitoring disease status based on time-series data modeling
Cancela et al. A comprehensive motor symptom monitoring and management system: the bradykinesia case
Tracy et al. Separating bedtime rest from activity using waist or wrist-worn accelerometers in youth
Sotirakis et al. Identification of motor progression in Parkinson’s disease using wearable sensors and machine learning
Crespo et al. Automatic identification of activity–rest periods based on actigraphy
Ihlen et al. The Discriminant Value of Phase‐Dependent Local Dynamic Stability of Daily Life Walking in Older Adult Community‐Dwelling Fallers and Nonfallers
JP6164678B2 (en) Detection apparatus, detection method, and detection program for supporting detection of signs of biological state transition based on network entropy
KR101947890B1 (en) Method and system for circadian rhythm calculation
WO2007046283A1 (en) Bioinformation acquiring device and bioinformation acquiring method
JP6198161B2 (en) Dynamic network biomarker detection apparatus, detection method, and detection program
JP2009519048A5 (en)
CN114999647A (en) Intelligent health monitoring method and system based on big data

Legal Events

Date Code Title Description
WITN Withdrawal due to no request for examination