Method for predicting depression by measuring emotional symptoms and device for performing same

Abstract

A method for predicting depression according to an embodiment of the present disclosure is performed by a computing device and comprises the steps of: generating a plurality of pieces of time series data of a plurality of emotional symptom measurements for a user during a specific period; generating emotional network data on the basis of the generated plurality of pieces of time series data; generating correlation network data by calculating correlation coefficients between each of the pieces of time series data in the generated emotional network data; and predicting depression symptoms of the user after the specific period by applying a machine-learning algorithm to the emotional network data or the correlation network data.

Description

Method for predicting depression by measuring emotional symptoms and device for performing the same

[0001] The present disclosure relates to a method for predicting depression by measuring emotional symptoms and a device for performing the same, and more specifically, to a method for predicting whether depression symptoms will worsen in the future by constructing a network using emotional symptom measurements for several days and applying a machine learning algorithm to the constructed network.

[0002] With the global increase in depression cases and its growing social impact, the diagnosis and early prediction of depression are becoming increasingly important. However, due to the nature of depression, many patients do not self-medicate. Clinical diagnosis currently relies on standardized diagnostic procedures based on questionnaire responses from patients who visit hospitals. Furthermore, clinically useful and reliable biomarkers remain unknown. Currently, research is focused on biological biomarkers measured in medical institutions, such as brain imaging, neurotransmitters, growth factors, and hormones, as well as digital biomarkers measured through digital devices, such as distance traveled, phone usage, and sleep status.

[0003] Biological biomarkers require a visit to a medical institution to be measured, limiting their potential for social application. Digital biomarkers, while potentially more comprehensive in the early prediction of depression, require large-scale data and provide information about an individual's lifestyle, hindering their widespread adoption. Therefore, a method is needed that eliminates the need for a visit to a medical institution or excessive demands on personal information, yet still achieves high accuracy in the diagnosis and early prediction of depression.

[0004] The technical problem to be solved by the embodiment of the present disclosure is to provide a method for predicting depression using voluntarily measured emotional symptom measurements.

[0005] In addition, a technical problem to be solved through the embodiment of the present disclosure is to provide a method for predicting depression by quantifying emotional symptoms from the usage history of digital devices such as smartphones.

[0006] A method for predicting depression according to an embodiment of the present disclosure may be performed by a computing device, and may include the steps of: generating a plurality of time series data measuring a plurality of emotional symptoms of a user during a specific period; generating emotional network data based on the generated plurality of time series data; generating correlation network data by calculating a correlation coefficient between each time series data of the generated emotional network data; and applying a machine learning algorithm to the emotional network data or the correlation network data to predict the depression symptoms of the user after the specific period.

[0007] In one embodiment, the plurality of emotional symptoms is characterized by including a plurality of positive emotional symptoms and a plurality of negative emotional symptoms.

[0008] In one embodiment, the step of generating the emotional network data may include the step of normalizing each of the generated time series data to the same scale, the step of interpolating missing values ​​of each of the generated time series data, and the step of generating matrix data including a plurality of time series data on which the normalization and interpolation have been performed as the emotional network data.

[0009] In one embodiment, the step of predicting the symptoms of depression may include a step of applying the machine learning algorithm to the correlation network data to predict whether the symptoms of depression are worsening, and a step of applying the machine learning algorithm to time series data of the correlation network data and the emotion network data to predict a change trend of emotion corresponding to the time series data.

[0010] In one embodiment, the machine learning algorithm is characterized as being a graph convolutional neural network (Graph CNN) algorithm.

[0011] In one embodiment, the step of generating the plurality of time series data is characterized in that the plurality of time series data are generated by receiving the results of measuring the plurality of emotional symptoms of the user from the user during the specific period.

[0012] In one embodiment, the step of generating the plurality of time series data is characterized in that the plurality of time series data is generated by quantifying the results of estimating the plurality of emotional symptoms from the user's digital device usage history during the specific period.

[0013] In one embodiment, the plurality of emotional symptoms are characterized in that they are estimated by extracting keywords representing emotions from text obtained from the digital device usage history.

[0014] According to another embodiment of the present disclosure, a computing device includes a processor and a memory storing instructions, which, when executed by the processor, cause the processor to perform the following operations: generating a plurality of time series data measuring a plurality of emotional symptoms of a user during a specific period of time; generating emotional network data based on the generated plurality of time series data; generating correlation network data by calculating a correlation coefficient between each time series data of the generated emotional network data; and applying a machine learning algorithm to the emotional network data or the correlation network data to predict the depression symptoms of the user after the specific period of time.

[0015] In one embodiment, the plurality of emotional symptoms is characterized by including a plurality of positive emotional symptoms and a plurality of negative emotional symptoms.

[0016] In one embodiment, the operation of generating the emotional network data may include an operation of normalizing each of the generated time series data to the same scale, an operation of interpolating missing values ​​of each of the generated time series data, and an operation of generating matrix data including a plurality of time series data on which the normalization and interpolation have been performed as the emotional network data.

[0017] In one embodiment, the operation of predicting the symptoms of depression may include an operation of applying the machine learning algorithm to the correlation network data to predict whether the symptoms of depression are getting worse, and an operation of applying the machine learning algorithm to time series data of the correlation network data and the emotion network data to predict a change trend of emotion corresponding to the time series data.

[0018] In one embodiment, the machine learning algorithm is characterized as being a graph convolutional neural network (Graph CNN) algorithm.

[0019] In one embodiment, the operation of generating the plurality of time series data is characterized by generating the plurality of time series data by receiving the results of measuring the plurality of emotional symptoms of the user from the user during the specific period.

[0020] In one embodiment, the operation of generating the plurality of time series data is characterized in that the plurality of time series data are generated by quantifying the results of estimating the plurality of emotional symptoms from the user's digital device usage history during the specific period.

[0021] In one embodiment, the plurality of emotional symptoms are characterized in that they are estimated by extracting keywords representing emotions from text obtained from the digital device usage history.

[0022] According to another embodiment of the present disclosure, a device for predicting depression may include a measurement module that generates a plurality of time series data measuring a plurality of emotional symptoms of a user during a specific period of time, a data generation module that generates emotional network data based on the generated plurality of time series data and calculates a correlation coefficient between each time series data of the generated emotional network data to generate correlation network data, and a prediction module that applies a machine learning algorithm to the emotional network data or the correlation network data to predict the depression symptoms of the user after the specific period of time.

[0023] FIG. 1 is a block diagram showing an exemplary configuration of a device for predicting depression according to an embodiment of the present disclosure.

[0024] FIG. 2 illustrates emotional network data and correlation network data according to an embodiment of the present disclosure.

[0025] FIG. 3 is a flowchart exemplarily showing a method for predicting depression according to an embodiment of the present disclosure.

[0026] Figure 4 is a flowchart specifically illustrating the steps for generating the emotional network data of Figure 3.

[0027] Figure 5 is a flowchart specifically showing the steps for predicting depression symptoms in Figure 3.

[0028] FIG. 6 is a flowchart exemplarily illustrating a method for predicting depression according to another embodiment of the present disclosure.

[0029] FIG. 7 is a block diagram showing the hardware configuration of a computing device for predicting depression according to an embodiment of the present disclosure.

[0030] Hereinafter, preferred embodiments of the present disclosure will be described in detail with reference to the attached drawings. The advantages and features of the present disclosure, and methods for achieving them, will become clearer with reference to the embodiments described in detail below together with the attached drawings. However, the present disclosure is not limited to the embodiments described below and may be implemented in various different forms. These embodiments are provided only to ensure that the present disclosure is complete and to fully inform those skilled in the art of the present disclosure of the scope of the invention, and the present disclosure is defined solely by the scope of the claims.

[0031] When designating components in each drawing, it should be noted that identical components are assigned identical reference numerals whenever possible, even if they appear on different drawings. Furthermore, when describing the present disclosure, detailed descriptions of related known components or functions are omitted if they are deemed to obscure the gist of the present disclosure.

[0032] Unless otherwise defined, all terms (including technical and scientific terms) used herein may be used in the same sense as commonly understood by those of ordinary skill in the art to which this disclosure pertains. Furthermore, terms defined in commonly used dictionaries are not to be interpreted ideally or excessively unless explicitly and specifically defined otherwise. The terminology used herein is for the purpose of describing embodiments and is not intended to limit the present invention. In this specification, singular forms also include plural forms, unless specifically stated otherwise.

[0033] Additionally, terms such as first, second, A, B, (a), (b), etc. may be used to describe components of the present disclosure. These terms are only intended to distinguish the components from other components, and the nature, order, or sequence of the components are not limited by the terms. When a component is described as being "connected," "coupled," or "connected" to another component, it should be understood that the component may be directly connected or connected to the other component, but another component may also be "connected," "coupled," or "connected" between each component.

[0034] The terms "comprises" and / or "comprising" as used in the specification do not exclude the presence or addition of one or more other components, steps, operations and / or elements.

[0035] FIG. 1 is a block diagram illustrating an exemplary configuration of a device (100) for predicting depression according to an embodiment of the present disclosure. Hereinafter, the device (100) for predicting depression will be referred to as a depression prediction device (100). The depression prediction device (100) generates emotional network data based on an individual's voluntary emotional symptom measurements for several days, and applies a machine learning algorithm to the generated emotional network data to predict whether or not depression symptoms will worsen in the future.

[0036] Referring to FIG. 1, the depression prediction device (100) may include a measurement module (110), a data generation module (120), and a prediction module (130). Meanwhile, the components (modules) of the depression prediction device (100) illustrated in FIG. 1 represent functionally distinct functional elements, and it should be noted that at least one component (module) may be implemented in a form that is integrated with each other in an actual physical environment.

[0037] The measurement module (110) can generate time-series data measuring an individual's emotional symptoms over a specific period. To this end, the measurement module (110) can provide a series of interfaces to the user that allow the user to voluntarily measure the emotional symptoms (i.e., input the individual's measurement results). For example, emotional symptoms can be categorized into positive emotional symptoms (e.g., cheerfulness, comfort, joy, etc.) and negative emotional symptoms (e.g., sadness, anger, fear, etc.), and each symptom can be measured with a numerical value (e.g., a Likert scale from 1 to 7). Alternatively, emotional symptoms can be measured according to valence, which indicates the degree of positivity or negativity of the emotion, and arousal, which indicates the degree of excitement due to the emotion, according to Russell's prototype emotion model. For the convenience of the following description, it is assumed that the measurement module (110) according to the embodiment of the present disclosure measures positive emotional symptoms and negative emotional symptoms; however, the present disclosure is not limited thereto, and the measurement module (110) may also measure an individual's emotional symptoms through various questionnaires that serve as a scale for diagnosing depression.

[0038] According to an embodiment, the measurement module (110) may quantify emotional symptoms based on an individual's digital device usage history. For example, the measurement module (110) may acquire text such as an individual's search history, the content of documents read on the web, the content of posts written on social media, and the content of posts liked on social media, and may extract keywords related to emotions from the acquired text to estimate and quantify emotional symptoms. The quantified emotional symptoms may likewise be generated as time-series data. The time-series data of emotional symptom measurements generated by the measurement module (110) may be provided to the data generation module (120).

[0039] The data generation module (120) generates emotional network data based on time-series data provided by the measurement module (110), and can generate correlation network data indicating correlations between each emotional symptom. Depending on the embodiment, the time-series data may be values ​​voluntarily measured by an individual, or values ​​estimated based on digital device usage history. For example, the emotional network data may be a matrix in which each row represents time-series data generated for each emotional symptom, and the correlation network data may be a symmetric matrix in which each element represents a correlation coefficient between time-series data. Specific examples of the emotional network data and the correlation network data are described in more detail with reference to FIG. 2.

[0040] When generating emotional network data, the data generation module (120) may perform preprocessing, such as normalizing each time series data to the same scale and interpolating missing values ​​of each time series data when the scales of each of the multiple time series data composed of emotional symptom measurement values ​​are different. The emotional network data and correlation network data generated through the data generation module (120) may be provided to the prediction module (130).

[0041] The prediction module (130) can predict future depression symptoms by applying a machine learning algorithm to the emotional network data or correlation network data provided from the data generation module (120). To this end, the prediction module (130) can use a machine learning model trained to predict depression symptoms through a large amount of emotional symptom measurements acquired in clinical settings. For example, the prediction module (130) can predict whether a specific individual will experience future worsening of depression symptoms by applying a graph convolutional neural network (Graph CNN) algorithm to the emotional network data and correlation network data (i.e., inputting the network data into a trained graph convolutional neural network). However, the present disclosure is not limited thereto, and the prediction module (130) can predict depression symptoms using various machine learning algorithms other than the graph convolutional neural network.

[0042] Specifically, the prediction module (130) can predict whether future depression symptoms will worsen by applying a machine learning algorithm to correlation network data. Since multiple emotional symptoms are complexly involved in each symptom of depression, the worsening of depression symptoms can be predicted through correlation network data that includes correlation coefficients between each emotional symptom. In addition, the prediction module (130) can also predict the change trend of a specific emotional symptom that may affect the worsening of depression symptoms by using the correlation network together with the measurement value of a specific emotional symptom included in the emotion network data. In this case, the measurement value of the emotional symptom can function as a feature that can be used in the machine learning algorithm. The prediction result through the prediction module (130) can be displayed to the user as a graphical interface.

[0043] Furthermore, based on the results of the predicted specific emotional symptoms, the prediction module (130) may provide feedback to the user to change the emotional symptoms in a way that alleviates the symptoms of depression. For example, if the user's positive emotional symptoms are predicted to weaken, the prediction module (130) may provide feedback to strengthen the user's positive emotional symptoms, and if the user's negative emotional symptoms are predicted to intensify, the prediction module (130) may provide feedback to alleviate the user's negative emotional symptoms. For example, the feedback may include playing music or videos frequently listened to by the user, recommending content the user likes, or presenting photos the user likes, based on the user's digital device usage history.

[0044] The depression prediction device (100) can be implemented in a computing device equipped with a computing means and a communication means, such as a smartphone, a desktop, a laptop, etc. For example, the depression prediction device (100) can be implemented as an application on a smartphone, measure the emotional symptoms of a smartphone user for several days, generate emotional network data and correlation network data based on the emotional symptoms measured for several days, and apply a machine learning algorithm to the generated network data to predict the depression symptoms of the smartphone user.

[0045] In addition, the depression prediction device (100) can communicate with an external server (e.g., a server storing clinical data related to depression) and all types of wired / wireless networks, such as a local area network (LAN), a wide area network (WAN), a mobile radio communication network, and Wibro (Wireless Broadband Internet), if necessary.

[0046] FIG. 2 illustrates emotional network data and correlation network data according to an embodiment of the present disclosure. The following description will be made with reference to FIG. 1, along with FIG. 2.

[0047] Let's assume that the measurement module (110) measured cheerfulness, comfort, and joy as positive emotional symptoms, and sadness, anger, and fear as negative emotional symptoms for 5 days (10). In this case, the data generation module (120) can generate emotional network data (20) in the form of a matrix of 6 rows and 5 columns using the time series data of each emotional symptom. Each row of the emotional network data (20) represents six different types of emotional symptoms, and each column represents a time point from the 1st to the 5th day. In addition, the data generation module (120) can generate correlation network data (30) in the form of a matrix of 6 rows and 6 columns based on the emotional network data (20). For example, in the correlation network data (30), x 12 can represent the correlation coefficient between the time series data corresponding to cheerfulness and the time series data corresponding to comfort, and x 42 can represent the correlation coefficient between the time series data corresponding to sadness and the time series data corresponding to comfort.

[0048] In addition, although not shown in FIG. 2, if there are missing values ​​in the emotional symptom measurement values ​​(10) as described with reference to FIG. 1, they can be interpolated by the data generation module (120), and if the scales of the measurement values ​​(10) are different, they can be normalized by the data generation module (120). The emotional network data (20) and the correlation network data (30) of FIG. 2 can be used to perform the depression symptom prediction operation of the prediction module (130).

[0049] FIG. 3 is a flowchart exemplifying a method for predicting depression according to an embodiment of the present disclosure. For reference, FIG. 3 illustrates steps / operations performed by the depression prediction device (100) of FIG. 1. Therefore, in the following descriptions, if the subject of a specific step / operation is omitted, it can be understood that the step / operation is performed by the depression prediction device (100) of FIG. 1. The following description will be made with reference to FIG. 1 along with FIG. 3.

[0050] In step S100, the measurement module (110) of the depression prediction device (100) may generate multiple time-series data measuring multiple emotional symptoms of the user over a specific period of time. For example, the measurement module (110) may receive the results of measuring multiple emotional symptoms from the user over a specific period of time. For example, the emotional symptoms may include multiple positive emotional symptoms and negative emotional symptoms, and the measurement results may be numerical values ​​expressed on an arbitrary scale (e.g., a Likert scale). In step S200, the data generation module (120) of the depression prediction device (100) may generate emotional network data based on the multiple time-series data generated by the measurement module (110). Each row of the emotional network data may represent time-series data corresponding to each emotional symptom. Step S200 will be described with reference to FIG. 4.

[0051] FIG. 4 is a flowchart specifically illustrating the step (S200) of generating emotional network data of FIG. 3. In step S210, the data generation module (120) can normalize each generated time-series data to the same scale. For example, the data generation module (120) can use min-max normalization or Z-score normalization as a normalization method, but the present disclosure is not limited thereto. Next, in step S220, the data generation module (120) can interpolate missing values ​​of each generated time-series data. Thereafter, in step S230, the data generation module (120) can generate matrix data including a plurality of time-series data on which normalization and interpolation have been performed as emotional network data.

[0052] Returning to Figure 3, in step S300, the data generation module (120) of the depression prediction device (100) can generate correlation network data by calculating correlation coefficients between each time series data of the generated emotional network data. For example, the correlation coefficient can be calculated as the Pearson correlation coefficient.

[0053] In step S400, the prediction module (130) of the depression prediction device (100) can apply a machine learning algorithm to emotional network data or correlation network data to predict the user's depression symptoms after a certain period of time. To this end, the prediction module (130) can use a trained machine learning model to receive emotional network data or correlation network data and predict depression symptoms. For example, the prediction module (130) can use a graph convolutional neural network algorithm as the machine learning algorithm. Step S200 will be described with reference to FIG. 5.

[0054] FIG. 5 is a flowchart specifically illustrating the step (S400) of predicting depression symptoms in FIG. 3. In step S410, the prediction module (130) can predict whether depression symptoms will worsen by applying a machine learning algorithm to the correlation network data. For example, since depression symptoms may be caused by a complex interaction of multiple emotional symptoms, depression symptoms may be predicted through the correlation of the correlation network data. Furthermore, in step S420, the prediction module (130) can predict the change trend of emotions corresponding to the time series data by applying a machine learning algorithm to the time series data of the correlation network data and the emotional network data. In this case, the time series data is used as a feature related to the machine learning algorithm.

[0055] FIG. 6 is a flowchart exemplifying a method for predicting depression according to another embodiment of the present disclosure. The following description will be made with reference to FIG. 1, along with FIG. 6, but any description overlapping with FIG. 3 will be omitted.

[0056] In step S100a, the measurement module (110) of the depression prediction device (100) can generate multiple time-series data by quantifying the results of estimating multiple emotional symptoms from the user's digital device usage history for a specific period of time. For example, the digital device usage history may include the user's search history, the contents of documents read on the web, the contents of posts written on SNS, the contents of posts liked on SNS, etc., and the estimation of emotional symptoms can be performed by extracting keywords related to emotions from texts obtained from such usage history. Steps S200 to S400 are performed in the same manner as described with reference to FIG. 3.

[0057] FIG. 7 is a block diagram showing the hardware configuration of a computing device (500) for predicting depression according to an embodiment of the present disclosure.

[0058] Referring to FIG. 7, a computing device (500) may include one or more processors (510), a bus (530), a communication interface (540), a memory (520) for loading a computer program executed by the processor (510), and a storage (550) for storing a computer program (560). However, only components related to the embodiment of the present disclosure are illustrated in FIG. 7. Therefore, a person skilled in the art to which the present disclosure pertains will appreciate that other general components may be included in addition to the components illustrated in FIG. 7. That is, the computing device (500) may further include various components in addition to the components illustrated in FIG. 7. In addition, in some cases, the computing device (500) may be configured in a form in which some of the components illustrated in FIG. 7 are omitted. Hereinafter, each component of the computing device (500) will be described.

[0059] The processor (510) can control the overall operation of each component of the computing device (500). The processor (510) can be configured to include at least one of a Central Processing Unit (CPU), a Micro Processor Unit (MPU), a Micro Controller Unit (MCU), a Graphics Processing Unit (GPU), or any other type of processor well known in the art of the present disclosure. In addition, the processor (510) can perform operations for at least one application or program for executing operations / methods according to embodiments of the present disclosure. The computing device (500) can include one or more processors.

[0060] Next, the memory (520) can store various data, commands, and / or information. The memory (520) can load a computer program (560) from the storage (550) to execute operations / methods according to embodiments of the present disclosure. The memory (520) may be implemented as a volatile memory such as RAM, but the present disclosure is not limited thereto.

[0061] Next, the bus (530) can provide communication functions between components of the computing device (500). The bus (530) can be implemented as various types of buses, such as an address bus, a data bus, and a control bus.

[0062] Next, the communication interface (540) can support wired and wireless Internet communication of the computing device (500). Furthermore, the communication interface (540) can also support various communication methods other than Internet communication. To this end, the communication interface (540) can be configured to include a communication module well known in the technical field of the present disclosure.

[0063] Next, the storage (550) can non-temporarily store one or more computer programs (560). The storage (550) can be configured to include non-volatile memory such as Read Only Memory (ROM), Erasable Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), flash memory, a hard disk, a removable disk, or any form of computer-readable recording medium well known in the art to which the present disclosure pertains.

[0064] Next, the computer program (560) may include one or more instructions that, when loaded into the memory (520), cause the processor (510) to perform operations / methods according to various embodiments of the present disclosure. That is, the processor (510) may perform operations / methods according to various embodiments of the present disclosure by executing the loaded one or more instructions.

[0065] For example, the computer program (560) may include instructions for generating multiple time series data measuring multiple emotional symptoms of a user over a specific period of time, generating emotional network data based on the generated multiple time series data, calculating correlation coefficients between each time series data of the generated emotional network data to generate correlation network data, and applying a machine learning algorithm to the emotional network data or the correlation data to predict depression symptoms. Alternatively, according to an embodiment, the computer program (560) may further include instructions for generating multiple time series data by quantifying the results of estimating multiple emotional symptoms from the user's digital device usage history over a specific period of time.

[0066] According to embodiments of the present disclosure, measuring biological biomarkers for diagnosing depression can eliminate the need for a visit to a medical institution, and the use of digital biomarkers requires little personal information, significantly improving utility and individual acceptance. Furthermore, the method for predicting depression according to embodiments of the present disclosure demonstrates a high level of predictive accuracy even without large-scale data, making it possible to quickly and easily predict the worsening of depressive symptoms and the trend of changes in specific emotions in advance.

[0067] In other words, according to embodiments of the present disclosure, it is possible to predict the likelihood of a potential patient, not yet diagnosed with depression, developing depression in the future. In other words, according to embodiments of the present disclosure, patients with depression who are unable to visit a hospital in person can easily diagnose the worsening of their symptoms on their own, making it possible to proactively screen and prevent future depression, rather than simply screening and treating current depressed patients.

[0068] Various embodiments of the present disclosure and the effects thereof have been described with reference to FIGS. 1 through 7. The effects of the technical concepts of the present disclosure are not limited to the effects described above, and other effects not mentioned will be readily apparent to those skilled in the art from the description below.

[0069] Although all components constituting the embodiments of the present disclosure have been described as being combined or operating in combination as one, the present disclosure is not necessarily limited to such embodiments. That is, within the scope of the present disclosure, all components may be selectively combined and operated in one or more combinations.

[0070] Although operations are depicted in the drawings in a particular order, this should not be understood to imply that the operations must be performed in the particular order depicted, or in any sequential order, or that all depicted operations must be performed to achieve the desired results. In certain circumstances, multitasking and parallel processing may be advantageous. Furthermore, the separation of the various components in the embodiments described above should not be understood to imply that such separation is absolutely necessary, and it should be understood that the program components and systems described may generally be integrated together into a single software product or packaged into multiple software products.

[0071] Although the embodiments of the present disclosure have been described with reference to the attached drawings, those skilled in the art will appreciate that the present disclosure can be implemented in other specific forms without altering the technical spirit or essential features thereof. Therefore, it should be understood that the embodiments described above are exemplary in all respects and not restrictive. The scope of protection of the present disclosure should be interpreted by the claims below, and all technical ideas within a scope equivalent thereto should be interpreted as being included within the scope of the rights of the present disclosure.

Claims

1. A method for predicting depression performed by a computing device, A step of generating multiple time series data measuring multiple emotional symptoms of a user over a specific period of time; A step of generating emotional network data based on the plurality of time series data generated above; A step of generating correlation network data by calculating the correlation coefficient between each time series data of the above-generated emotional network data; and A step of predicting the user's depression symptoms after the specific period of time by applying a machine learning algorithm to the emotional network data or the correlation network data, A method for predicting depression.

2. In paragraph 1, The above multiple emotional symptoms are characterized by including multiple positive emotional symptoms and multiple negative emotional symptoms. A method for predicting depression.

3. In paragraph 1, The steps for generating the above emotional network data are: A step of normalizing each of the above generated time series data to the same scale; A step of interpolating missing values ​​of each of the above-generated time series data; and A step of generating matrix data including a plurality of time series data on which the above normalization and interpolation have been performed as the emotional network data, A method for predicting depression.

4. In paragraph 1, The steps to predict the above depression symptoms are: A step of applying the machine learning algorithm to the above-mentioned correlation network data to predict whether depression symptoms are worsening; and A step of applying the machine learning algorithm to the time series data of the correlation network data and the emotion network data to predict the change trend of the emotion corresponding to the time series data, A method for predicting depression.

5. In paragraph 1, The above machine learning algorithm is characterized by being a graph convolutional neural network (Graph CNN) algorithm. A method for predicting depression.

6. In paragraph 1, The step of generating the plurality of time series data is characterized in that the plurality of time series data is generated by receiving the results of measuring the plurality of emotional symptoms of the user from the user during the specific period. A method for predicting depression.

7. In paragraph 1, The step of generating the plurality of time series data is characterized in that the plurality of time series data is generated by quantifying the results of estimating the plurality of emotional symptoms from the user's digital device usage history during the specific period. A method for predicting depression.

8. In paragraph 6, The above multiple emotional symptoms are characterized in that they are estimated by extracting keywords representing emotions from text obtained from the digital device usage history. A method for predicting depression.

9. Processor; and Contains memory that stores instructions, The above instructions, when executed by the processor, cause the processor to: An action to generate multiple time series data measuring multiple emotional symptoms of a user over a specific period of time; An operation of generating emotional network data based on the plurality of time series data generated above; An operation of generating correlation network data by calculating a correlation coefficient between each time series data of the above-mentioned generated emotional network data; and Applying a machine learning algorithm to the above emotional network data or the above correlation network data to perform an action of predicting the user's depression symptoms after the above specific period of time. Computing device.

10. In paragraph 9, The above multiple emotional symptoms are characterized by including multiple positive emotional symptoms and multiple negative emotional symptoms. Computing device.

11. In paragraph 9, The actions to generate the above emotional network data are: An operation to normalize each of the above generated time series data to the same scale; An operation of interpolating missing values ​​of each time series data generated above; and An operation of generating matrix data including a plurality of time series data on which the above normalization and interpolation have been performed as the emotional network data, Computing device.

12. In paragraph 9, Behaviors that predict the above depression symptoms are: An operation of predicting whether depression symptoms will worsen by applying the machine learning algorithm to the above-mentioned correlation network data; and An operation of applying the machine learning algorithm to the time series data of the correlation network data and the emotion network data to predict a change trend of emotion corresponding to the time series data, Computing device.

13. In paragraph 9, The above machine learning algorithm is characterized by being a graph convolutional neural network (Graph CNN) algorithm. Computing device.

14. In paragraph 9, The operation of generating the plurality of time series data is characterized in that it generates the plurality of time series data by receiving the results of measuring the plurality of emotional symptoms of the user from the user during the specific period. Computing device.

15. In paragraph 9, The operation of generating the plurality of time series data is characterized in that the plurality of time series data is generated by quantifying the results of estimating the plurality of emotional symptoms from the user's digital device usage history during the specific period. Computing device.

16. In paragraph 15, The above multiple emotional symptoms are characterized in that they are estimated by extracting keywords representing emotions from text obtained from the digital device usage history. Computing device.

17. A measurement module that generates multiple time series data measuring multiple emotional symptoms of a user over a specific period of time; A data generation module that generates emotional network data based on the plurality of generated time series data and calculates correlation coefficients between each time series data of the generated emotional network data to generate correlation network data; and A prediction module that applies a machine learning algorithm to the emotional network data or the correlation network data to predict the user's depression symptoms after the specific period of time, A device for predicting depression.

Images