KR20190030876A - Method for prediting health risk - Google Patents

Abstract

Disclosed is a computer program stored in a computer-readable storage medium having instructions encoded in accordance with one embodiment of the present disclosure. When executed by one or more processors of a computer system, the computer program causes the one or more processors to perform the following steps for indexing. The steps comprise: receiving biometric data including one or more items and change data different from the biometric data; inputting the biometric data and change data to a trained neural network and computing the biometric data and the change data using the trained neural network; generating first health risk information associated with the biometric data based on an output of the trained neural network and generating second health risk information associated with the change data; and generating a correlation between a data change amount of the biometric data and the change data and a health risk information change amount of the first health risk information and the second health risk information.

Description

{METHOD FOR PREDITING HEALTH RISK}

This disclosure relates to artificial intelligence techniques, and more specifically to predictions utilizing artificial intelligence techniques.

In recent years, various diseases such as cancer, hypertension, diabetes, and the like have been rapidly increasing due to aging and changes in lifestyle and various diseases that require continuous management. As the interest in health increases, Is increasing continuously.

Among these diseases, cardiovascular disease is one of the major causes of death, both in developing and developed countries. Globally, about 170 million people died in 2012, accounting for 31% of all deaths. Korea also has 50803 people as of 2013, accounting for about 20% of all deaths. Cardiovascular disease has such high mortality rates, but about 80% of all cardiovascular deaths can be prevented through blood pressure, blood sugar, cholesterol control and lifestyle changes. Thus, cardiovascular disease is highly predictive of disease outbreaks, as the accuracy of predictions for outbreaks is proportional to the mortality rate from disease.

In addition, it is known that genetic factors such as family history are important for diseases such as cancer, but it is also known to be caused by the life patterns such as the strains. However, it is difficult to precisely predict the stress medically, and there is a problem that the doctor can not accurately recognize the influence of the onset of the individual.

Thus, there may be a need in the art for a solution that can combine the causes of known diseases to predict disease outbreaks, and provide insight to the physician about the etiology of the disease.

Korean patent KR10-1744775 discloses a health management system that presents a mission for health care.

This disclosure is based on the background art described above and is intended to provide a solution for health care.

A computer program stored in a computer-readable storage medium having encoded instructions in accordance with one embodiment of the present disclosure for realizing the above-described subject matter is disclosed. Wherein the computer program causes the one or more processors to perform the following steps for indexing, when executed by one or more processors of a computer system, the steps comprising: Receiving biometric data and change data different from the biometric data; Inputting the biometric data and change data to a trained neural network, and computing the biometric data and the change data using the trained neural network; Generating first health risk information for the biometric data based on the output of the trained neural network and generating second health risk information for the change data; And generating a correlation between the data change amount of the biometric data and the change data, and the health risk information change amount of the first health risk information and the second health risk information.

Alternatively, the biometric data may include at least one of biometric data, demographic information, death related information, health assurance type information, socioeconomic level information, disability registration information, physical examination information, history information, family history information, And mental examination information may be included as the item.

Alternatively, extracting a highly related item that is highly related to the first health risk information among at least one item of the biometric data; And generating risk avoiding information for changing the highly correlated item.

Alternatively, the first health risk information may include at least one of probability of onset of cardiovascular disease, probability of onset of cancer, and probability of onset of adult disease.

Alternatively, the method may further include determining threshold change data in which a ratio of the data change amount and the health risk information change amount is equal to or greater than a predetermined threshold value, based on the association.

A computing device for predicting health risk incidence according to one embodiment of the disclosure is disclosed. The computing device may include one or more processors; And a memory for storing instructions executable by the one or more processors, wherein the one or more processors are configured to receive biometric data including one or more items and change data in which at least one item of the biometric data is modified, The biometric data and the change data are input to a trained neural network, the biometric data and the change data are calculated by using the trained neural network, and based on the output of the trained neural network, 1 health risk information, generates second health risk information on the change data, and generates health risk information on the basis of the data change amount of the biometric data, the change data, the health of the first health risk information, And generates a correlation of the amount of change in the risk information.

A method for predicting a health risk occurrence performed in one or more processors of a computing device is disclosed in accordance with one embodiment of the present disclosure. The method comprising: receiving biometric data including one or more items and change data in which at least one item of the biometric data is changed; Inputting the biometric data and change data to a trained neural network, and computing the biometric data and the change data using the trained neural network; Generating first health risk information for the biometric data based on the output of the trained neural network and generating second health risk information for the change data; And generating a correlation between the data change amount of the biometric data and the change data, and the health risk information change amount of the first health risk information and the second health risk information.

A computer program stored on a computer-readable storage medium comprising encoded instructions in accordance with another embodiment of the present disclosure is disclosed. The computer program, when executed by one or more processors of a computer system, causes the one or more processors to perform the following steps for indexing: obtaining biometrics data; Grouping the biometrics data by objects based on object classification information included in the biometrics data to generate samples for each object; Generating labeled learning data by matching the subject-specific sample with the clinical data; And generating a neural network model for training the neural network using the labeled learning data to output health risk information based on the biometric data.

Alternatively, the neural network may comprise a recurrent neural network (RNN), and the RNN may comprise a long short-term memory (LSTM).

Alternatively, the generating of the neural network model may include inputting the sample to the neural network, and calculating an error by comparing the output data calculated by the neural network with the medical care data; Adjusting a weight of the neural network based on the error in a backpropagation manner; Determining whether to stop learning using verification data if learning of the neural network is performed over a predetermined epoch; And testing the performance of the neural network model using the test data and determining whether to activate the neural network based on the performance.

Alternatively, the labeling includes whether or not the disease included in the medical care data, the name of the disease, and the time of occurrence, the learning data includes many years of biometric data, Or positive samples.

Alternatively, the step of generating the artificial positive sample by dividing the bio-measured data of the multi-year period with the time point of onset of the positive sample as the end point of the multi-year bio-measured data, And generating the artificial voice samples by dividing the biometric data of the multi-year sample of the voice sample so as to have an arbitrary start time and end time.

Alternatively, the method may further include the step of normalizing the measurement values included in the biometric data and setting a missing value included in the biometric data as an average value of the whole.

Alternatively,

The step of acquiring the biometric data may include acquiring the biometric data from at least one of an electronic health record (EHR), an electronic medical record (EMR), and a health examination DB from a hospital server and a government server And a step of acquiring.

A computing device for predicting health risk incidence according to another embodiment of the present disclosure is disclosed. The computing device may include one or more processors; And a memory for storing instructions executable by the one or more processors, wherein the one or more processors are configured to acquire biometrics data, and to store the biometrics data on a per-individual basis based on subject classification information included in the biometrics data Grouping and generating sample-by-object, generating learning data labeled by matching the target-specific sample with the medical care data, and training the neural network using the labeled learning data to output health risk information based on the biometric data To generate a neural network model.

A method for predicting a health risk occurrence performed in one or more processors of a computing device is disclosed in accordance with another embodiment of the present disclosure. The method includes obtaining biometrics data; Grouping the biometric data by individuals based on subject classification information included in the biometric data to generate a subject-specific sample; Generating labeled learning data by matching the subject-specific sample with the clinical data; And generating a neural network model for training the neural network using the labeled learning data to output health risk information based on the biometric data.

The present disclosure can provide a solution for health care.

1 is a block diagram of a computing device according to an embodiment of the present disclosure;
2 is a block diagram of a neural network according to an embodiment of the present disclosure.
3 is an exemplary user interface of a health risk prediction system according to one embodiment of the present disclosure;
4 is a flowchart of a health risk prediction method according to one embodiment of the present disclosure.
5 is a flowchart of a health risk prediction method according to another embodiment of the present disclosure.
Figure 6 is a simplified, general schematic diagram of an exemplary computing environment in which embodiments of the present disclosure may be implemented.

Various embodiments are now described with reference to the drawings. In this specification, various explanations are given in order to provide an understanding of the present disclosure. It will be apparent, however, that such embodiments may be practiced without these specific details. In other instances, well-known structures and devices are provided in block diagram form in order to facilitate describing the embodiments.

The terms "component," "module," system, "and the like, as used herein, refer to a computer-related entity, hardware, firmware, software, combination of software and hardware, or execution of software. For example, a component may be, but is not limited to, a process, a processor, an object, an execution thread, a program, and / or a computer running on a processor. For example, both an application running on a computing device and a computing device may be a component. One or more components may reside within a processor and / or thread of execution, one component may be localized within one computer, or it may be distributed between two or more computers. Further, such components may execute from various computer readable media having various data structures stored therein. The components may communicate, for example, via a network (e.g., a local system, data and / or signals from one component interacting with another component in a distributed system, and / or signals with one or more data packets) And / or < / RTI > transmitted data).

In addition, the term " comprises "and / or" comprising " means that the features and / or components are present, but excludes the presence or addition of one or more other features, components, and / It should be understood that it does not. Also, unless the context clearly dictates otherwise or to the contrary, the singular forms in this specification and claims should generally be construed to mean "one or more. &Quot;

The description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the scope of the present disclosure. Thus, the present disclosure should not be construed as limited to the embodiments set forth herein, but is to be accorded the widest scope consistent with the principles and novel features presented herein.

1 is a block diagram of a computing device according to an embodiment of the present disclosure;

The computing device 100 in accordance with one embodiment of the present disclosure may include a processor 110, a memory 130, and a network module 150. The block diagram shown in FIG. 1 is a simplified representation of a computing device, and the present disclosure is not so limited and may include additional components necessary for driving.

The processor 110 may be one or more of a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), a tensor processing unit (TPU) ), Etc., and a processor for deep-running. The processor 110 may read a computer program stored in the memory 130 to perform a health risk prediction method according to one embodiment of the present disclosure. In accordance with one embodiment of the present disclosure, the processor 110 may perform calculations for training of a neural network. The processor 110 is a processor for processing the input data for learning in deep learning (DN), extracting features from the input data, calculating errors, and updating the weight of the neural network using backpropagation. You can perform calculations for learning. Further, a computer program executed on a computing device according to an embodiment of the present disclosure may be a CPU, a GPGPU, or a TPU executable program.

The memory 130 may store a computer program for performing the health risk occurrence prediction method according to one embodiment of the present disclosure, and the stored computer program may be read and driven by the processor 110. [

The network module 150 can transmit and receive data and the like for performing a health risk occurrence prediction method according to an embodiment of the present disclosure with other computing devices, servers, and the like. The network module 150 can transmit and receive data required for embodiments of the present disclosure such as biometric data, change data, and the like to other computing devices, servers, and the like. For example, the network module 150 may receive an electronic health record (EHR) from a hospital server, a government server, and may also receive an electronic medical record (EMR) from a hospital server, can do. In addition, the network module 150 can receive a user's health check record or the like from a user terminal (not shown). The foregoing description is only illustrative and the present disclosure is not limited thereto.

2 is a block diagram of a neural network according to an embodiment of the present disclosure.

Throughout this specification, neural networks, network functions, and neural networks can be used interchangeably. A neural network may consist of a collection of interconnected computational units, which may be referred to generally as " nodes ". These " nodes " may also be referred to as " neurons ". The neural network is configured to include at least one or more nodes. The nodes (or neurons) that make up the neural networks may be interconnected by one or more " links ".

Within a neural network, one or more nodes connected through a link may form a relationship of input and output nodes relatively. The concepts of the input node and the output node are relative, and any node in the output node relationship with respect to one node can be in the input node relationship with the other node, and vice versa. As described above, the input node-to-output node relationship can be generated around the link. One or more output nodes may be connected to one input node via a link, and vice versa.

In an input node and an output node relationship connected through one link, the output node can be determined based on data input to the input node. Here, a node interconnecting the input node and the output node may have a weight. The weights may be variable and may be varied by the user or algorithm to perform the desired function of the neural network. For example, if one or more input nodes to one output node are interconnected by respective links, then the output node is set to the values input to the input nodes associated with the output node and to the corresponding links to the respective input nodes The output node value can be determined based on the weight.

As described above, the neural network is interconnected through one or more nodes over one or more links to form an input node and an output node relationship within the neural network. The nature of the neural network can be determined according to the number of nodes and links in the neural network, the association between the nodes and links, and the value of the weight assigned to each of the links. For example, if there are the same number of nodes and links, and there are two neural networks with different weight values between the links, then the two neural networks may be perceived as being different from each other.

As shown in FIG. 1, the neural network may be configured to include one or more nodes. Some of the nodes that make up the neural network can construct a layer based on the distances from the original input node. For example, n layers can be configured. The distance from the original input node may be defined by the minimum number of links that must go through to reach that node from the original input node. However, the definition of such a layer is arbitrary for explanation, and the degree of the layer in the neural network can be defined in a manner different from that described above. For example, the layer of nodes may be defined by distance from the final output node.

The first input node may refer to one or more nodes in the neural network to which data is directly input without going through a link in relation to other nodes. Or may refer to nodes in a neural network that do not have other input nodes connected by a link in the relationship between the nodes based on the link. Similarly, the final output node may refer to one or more nodes that do not have an output node, in relation to other ones of the nodes in the neural network. Also, the hidden node may refer to nodes constituting a neural network other than the first input node and the last output node. 1, the output node is omitted. The neural network according to an embodiment of the present disclosure may be a neural network in which the node of the input layer may be larger than the node of the hidden layer closer to the output layer and the number of nodes decreases as the input layer progresses to the hidden layer.

In one embodiment of the present disclosure, the neural network may include a recurrent neural network (RNN) to enable processing of time series data. Also, in one embodiment of the present disclosure, the neural network of the present disclosure may be implemented as a long short-term memory (LSTM) to prevent performance degradation due to a long range dependency vanishing gradient, . ≪ / RTI >

The RNN may be an artificial neural network that recognizes a pattern from data having the form of a sequence, such as a text, a gene, a handwriting, a voice signal, data sensed by a sensor, a stock price, and biometric data. RNN can process images as well as time series. Unlike normal feedforward neural networks, RNNs can be the output of the hidden layer again as input to the same hidden layer. The RNN is a neural network having a memory capability by calculating the input data currently input and the data received in the past simultaneously, and having a feedback structure. Thus, the RNN can be trained to interpret the current data according to the meaning of the previous data in time series data. One of the RNNs, called LSTMs, is also called long-term memory networks and can learn long-term dependencies. In one embodiment of the present disclosure, the neural network may include any neural network capable of processing time series data, such as a deep gated RNN, a clockwork RNN, etc., as well as an LSTM that is one of the RNNs .

The neural network 200 of FIG. 2 may include one or more hidden layers. The hidden node of the hidden layer can input the output of the previous layer and the output of the neighboring node. The number of hidden nodes for each hidden layer may be the same or different. The number of nodes of the input layer may be determined based on the number of data fields of the input data and may be equal to or different from the number of hidden nodes. The input data input to the input layer can be computed by the hidden node of the hidden layer and output by the fully connected layer (FCL), which is the output layer.

Hereinafter, a health risk occurrence prediction method according to an embodiment of the present disclosure will be described.

A method for generating a neural network model that predicts the occurrence of a health risk will be described in accordance with an embodiment of the present disclosure.

Processor 110 may obtain biometric data. Biometric data may include demographic information (eg, information on sex, age, residence area, etc.), death related information, health assurance type information, socioeconomic level information, disability registration information, physical examination information, And may include at least one of face image information, life history information, and mental examination information. The processor 110 can acquire biometric data by acquiring an electronic health record (EHR) and an electronic medical record (EMR) from a hospital server and a government server using the network module 150 have. For example, the biometric data may include a multi-year medical history record, may include information that can identify the subject, and may be anonymous or blindness information. For example, biometric data may include a qualification table and a health screening table of a health screening DB. The qualification table may include health insurance eligibility information of the health insurance subscribers and the medical benefit entitlement holder. For example, the eligibility table includes demographic information (including information on sex, age, and residence area), information related to death (including date of death, information related to death such as the cause of death), health insurance type information Information on health care benefits, etc.), socioeconomic level and other information (including information on income quintiles, disability registration information, etc.). Health check-up tables can include the main results of health screenings and questionnaire response data. For example, the health check-up table includes information related to test results such as general health checkups, life-time transition medical examination results, oral examination results, body measurements, blood tests, history information, family history, Test, depression, cognitive function test, and the like. In one embodiment of the present disclosure, the biometric data includes at least one of biometric time, sex, age, income quintile, severity of disability, type of disability, type of medical examination, body mass index, waist circumference, systolic blood pressure, diastolic blood pressure, , Triglyceride, HDL cholesterol, LDL cholesterol, hemoglobin, urine protein, serum creatinine, serum geoty, serum lipid, gamma-glycoprotein, hepatic disease, family history, family history of stroke, family history of heart disease, family history of hypertension, family history of diabetes History of smoking history, history of smoking history, smoking status, history of previous history of stroke, past history of heart disease, past history of hypertension, past history of diabetes, history of hyperlipidemia, past history of pulmonary tuberculosis, , Current smoking period, current average smoking amount per day, intense exercise for more than 20 minutes per week, intense exercise for more than 1 week 30 minutes, more than 30 minutes Cognitive function / comparison with peer group, cognitive function / remarks with 1 year ago, cognitive function / important cognitive function, cognitive function / other person's symptom recognition, cognitive function / And 59 main features such as the number of main main motions. The above-described biometric data is merely an example, and the present disclosure is not limited thereto.

The processor 110 may group the biometric data based on the target classification information included in the biometric data to generate a target-specific sample. The biometric data may include subject classification information. When the bio-measurement data is a multi-year health check record, the processor 110 may classify the health check information according to the target and generate the multi-year health check record as one object (i.e., individual). For example, when the biometric data is the health examination data of all the citizens from 2003 to 15, the processor 110 calculates the number of the examinees from 03 to 15 Year, and can generate a sample by subject. In the case of the health examination DB, information that does not include an individual is not included, but information that can distinguish who is each of the examination data is included, and the processor 110 uses the data to sort the biometric data by individual . The aforementioned subject-specific samples are merely illustrative, and the present disclosure is not limited thereto.

The processor 110 may generate the labeled training data by matching the subject-specific sample and the treatment data. In one embodiment of the present disclosure, the processor 110 may learn a neural network in a supervised learning manner. The medical care data may include the occurrence of the disease, the name of the disease, and the time of occurrence. The labeled learning data can be composed of input data and ordered pairs of correct answers. Here, the input data may include biometric data as described above, and the correct answer may include medical data. For example, the medical care data may include information of the medical care table of the health checkup DB. The medical treatment table may include medical use history and medical fee occurrence details on the medical care benefit cost statement. For example, the medical treatment table includes information on medical institution utilization, medical care costs, medical treatment subjects and medical illness information, examination, treatment, and other act payroll details, treatment material breakdown, treatment behavior breakdown, number code, detailed medical treatments, And the like. The learning data can be generated by matching the subject-specific sample and the clinical data. In the learning of the neural network, the subject-specific sample can be input to the input layer of the neural network, and the clinical data can be compared with the output of the neural network. The processor 110 may learn the neural network based on the computation result of the neural network for the subject-specific sample and the error of the clinical data (label). The learning data includes many years of biometrics data and can be classified as a negative sample or a positive sample depending on whether the disease has occurred or not. For example, neural network learning data for predicting the occurrence of cardiovascular disease may include a multi-year health check record and a medical record for the subject of the health checkup. The learning data may be classified into a positive sample in which the disease has occurred and a negative sample in which no disease has occurred based on the medical record (disease occurrence, name, occurrence time) of the subject. Positive samples may include many years of biometric data up to the time of disease occurrence. Positive samples may include years of biomedical data up to the time of disease occurrence, since events prior to the disease are likely to contain information that suggests a disease. In this case, the average of the data length of the positive sample (the follow-up length of biometric data included in the positive sample, for example, 5 years for biometric data for 5 years, 3 years for biometric data for 3 years, etc.) May be shorter than the average data length of the speech samples. Therefore, in order to prevent an error caused by a difference in the data length of the positive sample and the negative sample, the processor 110 sets the average of the data length of the voice sample to the data length of the positive sample, Only a part of the data can be selected and used.

When generating the training data, the processor 110 may normalize the measurement values contained in the biometric data. The processor 110 may reduce the computational complexity of neural network learning and normalize measurements to reduce errors due to unit and dimension differences in measurements. For example, the processor 110 may normalize the measurements included in the biometric data using a z-score. In addition, the processor 110 can set the missing value as the average value of the whole. For example, if the key of the patient a is missing and the key of the patient a is unknown in the result of the health examination for the patient a, the processor 110 may set the key of the patient a as the average value of the whole. In the normalized health examination results, the key of the patient a may be set to a normalized value of the average value of 0. [ Further, the processor 110 may set the missing value to the past record value using the past record of the subject, or to the average value of the past record. The above-described normalization method is merely an example, and the present disclosure is not limited thereto.

Processor 110 may generate artificial positive samples and artificial speech samples to increase the training data. The processor 110 can generate artificial positive samples by dividing bio-measurement data for many years with the time point of onset in the positive sample as the end point. For example, if the subject has a heart disease in 2007, the processor 110 determines a positive sample from 2002 to 2007 as the end point of the disease occurrence Artificial posi- tive samples from 2003 to 2007, and artificial posi- tive samples from 2005 to 2007). According to one embodiment of the present disclosure, by setting the end point of the artificial positive sample to the onset point, it is possible to eliminate the error that occurs when the end point of the artificial positive sample is set before the invention point (i.e., If you divide a sample into artificial samples from 2003 to 2006, you may experience errors that require forecasting of samples from 2006 to 2007). Processor 110 may divide multi-year biometric data of voice samples into arbitrary start and end times to generate artificial speech samples. For example, the processor 110 divides speech samples including biometric data from 2002 to 2010 into samples such as 2002 through 2004, 2006 through 2008, and generates artificial speech samples . In one embodiment of the present disclosure, the learning data to be learned by the neural network can not be arbitrarily increased as much as the biometric data of many years for realizing neural network learning. to be. As described above, according to one embodiment of the present disclosure, the processor 110 can generate artificial speech samples and artificial positive samples to increase learning data, thereby increasing the accuracy of neural network learning.

The processor 110 may generate a neural network model for training the neural network using the labeled learning data to output health risk information based on the biometric data. In one embodiment of the present disclosure, the trained neural network model can output health risk information based on input data when biometric data is input. The health risk information may include information on the probability of invention of any health risk that may occur in the living body, such as information on the invention of cardiovascular disease, information on the invention of cancer, and probability of invention of adult diseases. Of the onset of the onset of the disease.

The processor 110 may calculate an error by inputting a target-specific sample to the neural network to generate the neural network model and comparing the out-of-patient data calculated by the neural network. The processor 110 may adjust the weight of the neural network based on the error by a back propagation method. The processor 110 may set a drop out such that a portion of the output of the hidden node is not delivered to the next hidden node in order to prevent overfitting in the learning of the neural network. The processor 110 may set a learning rate to be high at the beginning of learning for learning efficiency, and may set the learning rate to be progressively lowered. The initial weight of each node of the neural network may be set based on a random or continuous even distribution.

After the processor 110 performs such learning over a predetermined epoch, the processor 110 may determine whether to stop the learning using the verification data. Where the predetermined epoch may be part of the overall learning target epoch. The processor 110 can test the performance of the generated neural network model using test data and determine whether to activate the neural network based on performance. The verification data may include many years of biometric data labeled. Test data may include many years of biometric data labeled. The processor 110 performs learning of the neural network using the training data. After the learning of the neural network is repeated over a predetermined epoch, the processor 110 can determine whether the learning effect of the neural network is equal to or higher than a predetermined level using the verification data. For example, when performing the learning with a target number of iterations of 100,000 times using one million learning data, the processor 110 performs an iterative learning of 10,000 times, which is a pre-determined epoch, and then uses 1000 verification data 10 times of iterative learning is performed, and if the change in the output of the neural network is less than or equal to the predetermined level during the 10 repetitions of learning, it is judged that the further learning is meaningless and the learning can be terminated. That is, the verification data can be used to determine completion of learning based on whether or not the effect of epilepsy learning in a repeated learning of the neural network is equal to or greater than a certain level. The above-described learning data, the number of verification data, and the number of repetition times are merely examples, and the present disclosure is not limited thereto.

Test data can be used to verify the performance of the neural network. The processor 110 may input test data to the neural network for which learning has been completed and measure the error to determine whether to activate the neural network based on whether the neural network is above a predetermined performance criterion. The processor 110 may verify the performance of the learned neural network using the test data to the neural network that has been already learned and may activate the neural network to be used in another application when the performance of the learned neural network is higher than a predetermined reference. In addition, the processor 110 may deactivate the neural network and discard the neural network if the performance of the learned neural network is below a predetermined threshold. For example, the processor 110 may determine the performance of the generated neural network model based on factors such as accuracy, precision, and recall. The performance evaluation criteria described above are merely examples, and the present disclosure is not limited thereto. According to an embodiment of the present disclosure, the processor 110 may independently learn each of the neural networks to generate a plurality of neural network models. Also, in one embodiment of the present disclosure, one or more neural networks may be used for health risk prediction and, when a plurality of neural networks are used, the outputs of the plurality of neural networks may be combined to generate health risk prediction information.

According to one embodiment of the present disclosure, a neural network model may be generated that predicts the occurrence of health risks based on multi-year health screening data.

A method for predicting the occurrence of a health risk using a trained neural network according to one embodiment of the present disclosure will be described.

The processor 110 may receive biometric data including one or more items and change data different from the biometric data. Biometric data include at least one of demographic information, death related information, health assurance type information, socioeconomic level information, disability registration information, physical examination information, history information, family history information, lifestyle information, One item may be included as the item. The change data includes data in which one or more items of the biometric data are changed. For example, if the biometric data is a 28-year-old male blood glucose of 80 mg / dL, the change data could be a 28-year-old male blood glucose of 95 mg / dL. Biometric data may be obtained from an external database or a user terminal. For example, referring to an exemplary user interface of a health risk prediction system according to one embodiment of the present disclosure of FIG. 3, the interface example of FIG. 3 may be an example of an interface displayed on a user terminal. 3, the user can input biometric data and change data to the biometric data input interface 310, and the processor 110 can receive the input biometric data and change data. In addition, the user can photograph the result of the health examination using the camera of the user terminal, and the computing device 100 according to an embodiment of the present disclosure may receive the image from the user terminal and acquire biometrics data.

The processor 110 may input biometric data and change data into a trained neural network, and may use the trained neural network to calculate biometric data and change data. The trained neural network may be a neural network trained by a method of generating a neural network model that predicts the occurrence of health risks according to one embodiment of the present disclosure described above.

The processor 110 may generate first health risk information for the biometric data based on the output of the trained neural network and generate second health risk information for the change data. The health risk information may include, for example, information related to the probability of occurrence of a health hazard occurring in the body such as information on the probability of occurrence of cardiovascular disease, probability of occurrence of cancer, and probability of occurrence of adult disease, and timing. For example, referring to FIG. 3, the processor 110 may generate health risk information based on biometric data and change data, and the computing device 100 may transmit the health risk information to the user terminal, And displayed on the interface 330.

The processor 110 may extract a highly correlated item that is highly correlated with the first health risk information of one or more items of the biometric data. The processor 110 may also generate risk avoidance information for changing the high association item. The processor 110 can extract the input items that are highly related to the output of the neural network by grasping the connection relation of the trained neural network and the connection weight of each node. The processor 110 may perform back projection on the input from the output of the neural network based on the input bio-measured data to determine an item of input data that plays a decisive role in the output value. For example, when the probability of occurrence of heart disease is high in the output of the neural network, the processor 110 may inversely calculate the output of the neural network to determine an item of input data that has a decisive influence on the occurrence of heart disease . For example, the entry in the input data that has had a decisive impact on the occurrence of the disease may be cholesterol as it is already known, or it may be the income quintile, which is generally considered to be of low relevance. The above-mentioned items are merely examples, and the present disclosure is not limited thereto. Thus, in accordance with one embodiment of the present disclosure, the processor 110 may perform operations on the input at the output of the neural network to extract critical items that cause the patient a predicted health risk, Can be used to extract other causes in which the relationship is learned.

Accordingly, the health risk prediction method according to one embodiment of the present disclosure can provide an insight into the causal relationship of diseases and diseases other than the cause of the disease generally known to the clinician.

 Processor 110 may generate risk avoidance information for changing the high association item. For example, if the processor 110 determines that a level of triglyceride above a certain level of the patient is highly related to the onset of heart disease using a trained neural network, then the processor 110 may And can generate lifestyle, eating habits, medical treatment, and drug information to reduce the triglycerides that can be provided to the patient. The above-mentioned high association items are only illustrative, and the present disclosure is not limited thereto.

The processor 110 may generate a correlation between the data change amount of the biometric data and the change data, and the health risk information change amount of the first health risk information and the second health risk information. For example, if the biometric data is for a 28-year-old male blood glucose of 80 mg / dL, the change data may be a 15 mg / dL increase in blood glucose for a 28-year-old male blood glucose of 95 mg / dL. At this time, the processor 110 generates a correlation between the data change amount of the biometrics data and the change data, the first health risk information, and the second health risk information, and detects a heart disease (Health risk information) of the occurrence probability of the health risk. The processor 110 can generate a correlation between an increased probability of occurrence of a heart disease and an increased blood sugar amount in the case where the blood glucose level is increased. For example, the processor 110 may generate a 5% increase in the probability of occurrence of heart disease if the blood glucose level is increased by 15 mg / dL based on the output of the trained neural network. The foregoing data and health risk information are illustrative only and the present disclosure is not so limited.

These associations may be generated through computations using the learned neural network and may be the same or different from common medical sense. In the case of common medical common sense, doctors can use this information to accurately grasp the linkage between factors and risk, and if they are different from common medical common sense, doctors can use them to discover new disease factors and provide new insights to physicians can do.

The processor 110 can determine the threshold change data whose ratio of the data change amount and the health risk information change amount is equal to or larger than the predetermined threshold based on this association. The threshold change data may include items and numerical values of the biometric data when the biometric data slightly affect only the health risk information. That is, the threshold change data may include biometric data having critical significance to health risk information. For example, a physician can use a trained neural network according to one embodiment of the present disclosure to observe changes in the output of neural networks (i.e., health risk information) by slightly changing the numerical values of some items in the patient's biometric data It is assumed that an experiment is performed. For example, if a physician increases the blood glucose level by 1 mg / dL and observes the probability of developing heart disease, if the probability of developing heart disease increases sharply when blood glucose reaches 100 mg / dL to 101 mg / dL (for example, The blood glucose level is increased more than the increase in the probability of occurrence of heart disease when the blood glucose level is increased from 99 mg / dL to 100 mg / dL), the blood glucose level 101 mg / dL may be the threshold change data for the patient. The foregoing description is only illustrative and the present disclosure is not limited thereto. In other words, the processor 110 can determine the threshold change data based on the change amount of the health risk information when the change amount of the health risk information is larger than the change of the other sections of the biometrics data. The processor 110 may also determine the item as the threshold change data if the change amount of the health risk information has a critical meaning even if the item including the change data is within the normal value. For example, if a physician or researcher changes the blood glucose within a normal range and tests for the risk of developing a disease, if the probability of a health risk is critical even if the blood sugar is within the normal range, It can be determined by the threshold change data. The foregoing description is only illustrative and the present disclosure is not limited thereto.

Experiments on causative factors require control over factors other than the causative factor of the invention, but biometric data are extremely difficult to obtain in terms of their characteristics. For example, it is almost impossible to obtain data with all other biometric values being the same and having different blood glucose values. Thus, using a trained neural network according to one embodiment of the present disclosure, a physician or researcher can experiment with the probability of onset by changing the value of the causative agent to be tested. .

4 is a flowchart of a health risk prediction method according to one embodiment of the present disclosure.

The computing device 100 may receive (410) biometric data that includes one or more items and change data that is different from the biometric data. Biometric data include at least one of demographic information, death related information, health assurance type information, socioeconomic level information, disability registration information, physical examination information, history information, family history information, lifestyle information, One item may be included as the item. The change data includes data in which one or more items of the biometric data are changed. For example, if the biometric data is a 28-year-old male blood glucose of 80 mg / dL, the change data could be a 28-year-old male blood glucose of 95 mg / dL. Biometric data may be obtained from an external database or a user terminal.

The computing device 100 may input biometric data and change data into the trained neural network, and may use the trained neural network to calculate biometric data and change data (420). The trained neural network may be a neural network trained by a method of generating a neural network model that predicts the occurrence of health risks according to one embodiment of the present disclosure described above.

The computing device 100 may generate first health risk information for biometric data based on the output of the trained neural network and generate second health risk information for the change data (430). The health risk information may include, for example, information related to the probability of occurrence of a health hazard occurring in the body such as information on the probability of occurrence of cardiovascular disease, probability of occurrence of cancer, and probability of occurrence of adult disease, and timing.

The computing device 100 may extract a highly correlated item that is highly correlated with the first health risk information of one or more items of the biometric data. The computing device 100 may also generate risk avoidance information for changing the high association item. The computing device 100 can extract input items having a high correlation with the output of the neural network by grasping the connection relation of the trained neural network and the connection weight of each node. The computing device 100 may perform back projection on the input from the output of the neural network based on the input bio-measured data to determine an item of input data that plays a decisive role in the output value. For example, when the computing device 100 predicts that the probability of occurrence of a heart disease is high in the output of a neural network, the computing device 100 calculates an output of the neural network inversely to determine an item of input data that has a decisive influence on the occurrence of a heart disease can do. For example, the entry in the input data that has had a decisive impact on the occurrence of the disease may be cholesterol as it is already known, or it may be the income quintile, which is generally considered to be of low relevance. The above-mentioned items are merely examples, and the present disclosure is not limited thereto. Thus, in accordance with one embodiment of the present disclosure, the computing device 100 may perform operations on the input at the output of the neural network to extract critical items that cause a predictable health risk to the patient, It is possible to extract other causes in which the relationship is learned by the neural network.

The computing device 100 may generate a correlation between the data change amount of the biometrics data and the change data, and the health risk information change amount of the first health risk information and the second health risk information (440). For example, if the biometric data is for a 28-year-old male blood glucose of 80 mg / dL, the change data may be a 15 mg / dL increase in blood glucose for a 28-year-old male blood glucose of 95 mg / dL. At this time, the processor 110 generates a correlation between the data change amount of the biometrics data and the change data, the first health risk information, and the second health risk information, and detects a heart disease (Health risk information) of the occurrence probability of the health risk. The computing device 100 may generate a correlation between an increased probability of occurrence of a heart disease and an increased blood sugar amount in the case where the blood glucose level is increased. For example, the computing device 100 may generate a 5% increase in the probability of developing heart disease if the blood glucose level is increased by 15 mg / dL based on the output of the trained neural network. The foregoing data and health risk information are illustrative only and the present disclosure is not so limited.

The computing device 100 may generate risk avoidance information for changing the highly correlated item. For example, if the computing device 100 determines that a level of triglyceride above a certain level of the patient is highly related to the onset of heart disease using the trained neural network, the computing device 100 may avoid And may generate lifestyle, dietary habits, medical treatment, and drug information to reduce the triglycerides that may be provided to the patient. The above-mentioned high association items are only illustrative, and the present disclosure is not limited thereto.

The computing device 100 may generate a correlation between the data change amount of the biometric data and the change data, and the health risk information change amount of the first health risk information and the second health risk information. For example, if the biometric data is for a 28-year-old male blood glucose of 80 mg / dL, the change data may be a 15 mg / dL increase in blood glucose for a 28-year-old male blood glucose of 95 mg / dL. At this time, the computing device 100 generates a correlation between the data change amount of the biometrics data and the change data, the first health risk information, and the second health risk information, and when the blood glucose level is increased by 15 mg / dL, It is possible to calculate a change amount of the occurrence probability (health risk information) of the disease. The computing device 100 may generate a correlation between an increased probability of occurrence of a heart disease and an increased blood sugar amount in the case where the blood glucose level is increased. For example, the processor 110 may generate a 5% increase in the probability of occurrence of heart disease if the blood glucose level is increased by 15 mg / dL based on the output of the trained neural network. The foregoing data and health risk information are illustrative only and the present disclosure is not so limited.

These associations may be generated through computations using the learned neural network and may be the same or different from common medical sense. In the case of common medical common sense, doctors can use this information to accurately grasp the linkage between factors and risk, and if they are different from common medical common sense, doctors can use them to discover new disease factors and provide new insights to physicians can do.

The computing device 100 can determine the threshold change data whose ratio of the data change amount to the health risk information change amount is equal to or greater than the predetermined threshold value based on this association. The threshold change data may include items and numerical values of the biometric data when the biometric data slightly affect only the health risk information. That is, the threshold change data may include biometric data having critical significance to health risk information. For example, a physician can use a trained neural network according to one embodiment of the present disclosure to observe changes in the output of neural networks (i.e., health risk information) by slightly changing the numerical values of some items in the patient's biometric data It is assumed that an experiment is performed. For example, if a physician increases the blood glucose level by 1 mg / dL and observes the probability of developing heart disease, if the probability of developing heart disease increases sharply when blood glucose reaches 100 mg / dL to 101 mg / dL (for example, The blood glucose level is increased more than the increase in the probability of occurrence of heart disease when the blood glucose level is increased from 99 mg / dL to 100 mg / dL), the blood glucose level 101 mg / dL may be the threshold change data for the patient. The foregoing description is only illustrative and the present disclosure is not limited thereto. That is, when the amount of change in the health risk information is larger than the change in the other sections of the biometric data, the computing device 100 can determine the threshold change data based on the change amount.

Experiments on causative factors require control over factors other than the causative factor of the invention, but biometric data are extremely difficult to obtain in terms of their characteristics. For example, it is almost impossible to obtain data with all other biometric values being the same and having different blood glucose values. Thus, using a trained neural network according to one embodiment of the present disclosure, a physician or researcher can experiment with the probability of onset by changing the value of the causative agent to be tested. .

5 is a flowchart of a health risk prediction method according to another embodiment of the present disclosure.

The computing device 100 may obtain biometric data (510). Biometric data may include demographic information (eg, information on sex, age, residence area, etc.), death related information, health assurance type information, socioeconomic level information, disability registration information, physical examination information, And may include at least one of face image information, life history information, and mental examination information. The computing device 100 may obtain biometric data by obtaining an electronic health record (EHR) and an electronic medical record (EMR) from a hospital server, a government server, and the like. For example, the biometric data may include a multi-year medical history record, may include information that can identify the subject, and may be anonymous or blindness information. For example, biometric data may include a qualification table and a health screening table of a health screening DB. The foregoing description is only illustrative and the present disclosure is not limited thereto.

The computing device 100 may group the biometric data based on the target classification information included in the biometric data to generate a target-specific sample (520). The biometric data may include subject classification information. When the bio-measurement data is a multi-year health check record, the computing device 100 can generate the multi-year health check record as one object (i.e., individual) by classifying the relevant health check information by object. For example, when the biometric data is the national health examination data from 2003 to 15, the computing device 100 calculates the data of the examinees based on information that can identify the examinees included in the data, It can be grouped into up to 15 years of screening data to generate sample by subject. In the case of the health examination DB, information that does not include an individual is not included, but includes information that can distinguish who is each of the examination data, and the computing device 100 uses the biometric data as individual Grouping can be performed. The aforementioned subject-specific samples are merely illustrative, and the present disclosure is not limited thereto.

The computing device 100 may generate the labeled learning data by matching the subject-specific sample with the clinical data (530). In one embodiment of the present disclosure, the computing device 100 may learn a neural network in a supervised learning manner. The medical care data may include the occurrence of the disease, the name of the disease, and the time of occurrence. The labeled learning data can be composed of input data and ordered pairs of correct answers.

When the computing device 100 generates training data, it may normalize the measurement values contained in the biometric data.

Computing device 100 may generate artificial positive samples and artificial speech samples to increase learning data. The computing device 100 can generate artificial positive samples by dividing bio-measurement data for many years with the time point of onset in the positive sample as the end point.

The computing device 100 may generate a neural network model for training neural networks using labeled learning data to output health risk information based on biometric data. In one embodiment of the present disclosure, the trained neural network model can output health risk information based on input data when biometric data is input. The health risk information may include information on the probability of invention of any health risk that may occur in the living body, such as information on the invention of cardiovascular disease, information on the invention of cancer, and probability of invention of adult diseases. Of the onset of the onset of the disease.

The computing device 100 may calculate an error by inputting a target-specific sample into a neural network to generate a neural network model, and comparing the out-of-patient data calculated by the neural network. The computing device 100 may adjust the weight of the neural network in a back propagation manner based on the error. The computing device 100 may set a drop out such that a portion of the output of the hidden node is not delivered to the next hidden node in order to prevent overfitting in the learning of the neural network. The computing device 100 may set the learning rate to be high at the beginning of learning and set the learning rate to be low for learning efficiency. The initial weight of each node of the neural network may be set based on a random or continuous even distribution.

After the computing device 100 performs such learning over a predetermined epoch, it can use the verification data to determine whether to stop learning. The computing device 100 may test the performance of the neural network model generated using the test data and determine whether to activate the neural network based on performance. The verification data may include many years of biometric data labeled. Test data may include many years of biometric data labeled. The computing device 100 may perform learning of the neural network using the training data and may determine whether the learning effect of the neural network is equal to or higher than a predetermined level by using the verification data after the learning of the neural network is repeated over a predetermined epoch. For example, when performing the learning with a target number of iterative learning times of 100,000 times using one million learning data, the computing device 100 performs an iterative learning of 10,000 times, which is a pre-determined epoch, 10 times of iterative learning is performed. If the change of the output of the neural network is less than or equal to the predetermined level during the ten iterations of learning, it is judged that further learning is meaningless and the learning can be completed. That is, the verification data can be used to determine completion of learning based on whether or not the effect of epilepsy learning in a repeated learning of the neural network is equal to or greater than a certain level. The above-described learning data, the number of verification data, and the number of repetition times are merely examples, and the present disclosure is not limited thereto.

Test data can be used to verify the performance of the neural network. The computing device 100 may input test data to the neural network that has completed the learning and measure the error to determine whether to activate the neural network based on whether the neural network is above a predetermined performance criterion. The computing device 100 may verify the performance of the learned neural network using the test data to the neural network that has completed the learning and activate the neural network to be used in another application when the performance of the learned neural network is higher than a predetermined reference. Also, the computing device 100 may deactivate and discard the neural network if the performance of the learned neural network is below a predetermined criterion. For example, the computing device 100 may determine the performance of the generated neural network model based on factors such as accuracy, precision, and recall. The performance evaluation criteria described above are merely examples, and the present disclosure is not limited thereto. According to an embodiment of the present disclosure, the processor 110 may independently learn each of the neural networks to generate a plurality of neural network models. Also, in one embodiment of the present disclosure, one or more neural networks may be used for health risk prediction and, when a plurality of neural networks are used, the outputs of the plurality of neural networks may be combined to generate health risk prediction information.

According to one embodiment of the present disclosure, a neural network model may be generated that predicts the occurrence of health risks based on multi-year health screening data.

Figure 6 is a simplified, general schematic diagram of an exemplary computing environment in which embodiments of the present disclosure may be implemented.

Although the present disclosure has been described above generally in terms of computer-executable instructions that may be executed on one or more computers, those skilled in the art will appreciate that the disclosure may be combined with other program modules and / will be.

Generally, program modules include routines, programs, components, data structures, etc. that perform particular tasks or implement particular abstract data types. Those skilled in the art will also appreciate that the methods of the present disclosure may be practiced with other computer systems, including single-processor or multiprocessor computer systems, minicomputers, mainframe computers, as well as personal computers, handheld computing devices, microprocessor-based or programmable consumer electronics, And may operate in conjunction with one or more associated devices).

The described embodiments of the present disclosure may also be practiced in distributed computing environments where certain tasks are performed by remote processing devices connected through a communications network. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.

Computers typically include a variety of computer readable media. Any medium accessible by a computer may be a computer-readable medium, which may include volatile and non-volatile media, transitory and non-transitory media, removable and non-removable media, Removable media. By way of example, and not limitation, computer readable media can comprise computer readable storage media and computer readable transmission media. Computer-readable storage media includes both volatile and non-volatile media, both temporary and non-volatile media, both removable and non-removable, implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data Media. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital video disk (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, Or any other medium which can be accessed by a computer and used to store the desired information.

Computer readable transmission media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism, It includes all information delivery media. The term modulated data signal refers to a signal that has one or more of its characteristics set or changed to encode information in the signal. By way of example, and not limitation, computer readable transmission media include wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared, or other wireless media. Combinations of any of the above described media are also intended to be included within the scope of computer readable transmission media.

There is shown an exemplary environment 1100 that implements various aspects of the present disclosure including a computer 1102 and a computer 1102 that includes a processing unit 1104, a system memory 1106, and a system bus 1108 do. The system bus 1108 couples system components, including but not limited to, system memory 1106 to the processing unit 1104. The processing unit 1104 may be any of a variety of commercially available processors. Dual processors and other multiprocessor architectures may also be used as the processing unit 1104.

The system bus 1108 may be any of several types of bus structures that may additionally be interconnected to a local bus using any of the memory bus, peripheral bus, and various commercial bus architectures. The system memory 1106 includes read only memory (ROM) 1110 and random access memory (RAM) The basic input / output system (BIOS) is stored in a non-volatile memory 1110, such as a ROM, EPROM, EEPROM or the like, which is a basic (non-volatile) memory device that aids in transferring information between components within the computer 1102 Routine. The RAM 1112 may also include a high speed RAM such as static RAM for caching data.

The computer 1102 may also be an internal hard disk drive (HDD) 1114 (e.g., EIDE, SATA) - this internal hard disk drive 1114 may also be configured for external use within a suitable chassis , A magnetic floppy disk drive (FDD) 1116 (e.g., for reading from or writing to a removable diskette 1118), and an optical disk drive 1120 (e.g., a CD-ROM For reading disc 1122 or reading from or writing to other high capacity optical media such as DVD). The hard disk drive 1114, magnetic disk drive 1116 and optical disk drive 1120 are connected to the system bus 1108 by a hard disk drive interface 1124, a magnetic disk drive interface 1126 and an optical drive interface 1128, respectively. . The interface 1124 for external drive implementation includes at least one or both of Universal Serial Bus (USB) and IEEE 1394 interface technologies.

These drives and their associated computer-readable media provide non-volatile storage of data, data structures, computer-executable instructions, and the like. In the case of computer 1102, the drives and media correspond to storing any data in a suitable digital format. While the above description of computer readable media refers to HDDs, removable magnetic disks, and removable optical media such as CDs or DVDs, those skilled in the art will appreciate that other types of storage devices, such as zip drives, magnetic cassettes, flash memory cards, Or the like may also be used in the exemplary operating environment and any such medium may include computer-executable instructions for carrying out the methods of the present disclosure.

A number of program modules may be stored in the drive and RAM 1112, including an operating system 1130, one or more application programs 1132, other program modules 1134, and program data 1136. All or a portion of the operating system, applications, modules, and / or data may also be cached in the RAM 1112. It will be appreciated that the disclosure may be implemented in a variety of commercially available operating systems or combinations of operating systems.

A user may enter commands and information into the computer 1102 via one or more wired / wireless input devices, such as a keyboard 1138 and a pointing device such as a mouse 1140. [ Other input devices (not shown) may include a microphone, IR remote control, joystick, game pad, stylus pen, touch screen, and so on. These and other input devices are often connected to the processing unit 1104 via an input device interface 1142 that is coupled to the system bus 1108, but may be a parallel port, an IEEE 1394 serial port, a game port, a USB port, an IR interface, ≪ / RTI > and so forth.

A monitor 1144 or other type of display device is also connected to the system bus 1108 via an interface, such as a video adapter 1146, In addition to the monitor 1144, the computer typically includes other peripheral output devices (not shown) such as speakers, printers,

Computer 1102 may operate in a networked environment using logical connections to one or more remote computers, such as remote computer (s) 1148, via wired and / or wireless communication. The remote computer (s) 1148 may be a workstation, a computing device computer, a router, a personal computer, a portable computer, a microprocessor-based entertainment device, a peer device or other conventional network node, But for the sake of simplicity, only memory storage device 1150 is shown. The logical connections depicted include a wired / wireless connection to a local area network (LAN) 1152 and / or a larger network, e.g., a wide area network (WAN) These LAN and WAN networking environments are commonplace in offices and corporations and facilitate enterprise-wide computer networks such as intranets, all of which can be connected to computer networks worldwide, for example the Internet.

When used in a LAN networking environment, the computer 1102 is connected to the local network 1152 via a wired and / or wireless communication network interface or adapter 1156. [ The adapter 1156 may facilitate wired or wireless communication to the LAN 1152 and the LAN 1152 also includes a wireless access point installed therein to communicate with the wireless adapter 1156. [ When used in a WAN networking environment, the computer 1102 may include a modem 1158, or may be connected to a communications computing device on the WAN 1154, or to establish communications over the WAN 1154 And other means. A modem 1158, which may be an internal or external and a wired or wireless device, is coupled to the system bus 1108 via a serial port interface 1142. In a networked environment, program modules described for the computer 1102, or portions thereof, may be stored in the remote memory / storage device 1150. It will be appreciated that the network connections shown are exemplary and other means of establishing a communication link between the computers may be used.

The computer 1102 may be any wireless device or entity that is deployed and operable in wireless communication, such as a printer, a scanner, a desktop and / or portable computer, a portable data assistant (PDA) Any equipment or place, and communication with the telephone. This includes at least Wi-Fi and Bluetooth wireless technology. Thus, the communication may be a predefined structure, such as in a conventional network, or simply an ad hoc communication between at least two devices.

Wi-Fi (Wireless Fidelity) allows you to connect to the Internet without wires. Wi-Fi is a wireless technology such as a cell phone that allows such devices, e.g., computers, to transmit and receive data indoors and outdoors, i. E. Anywhere within the coverage area of a base station. Wi-Fi networks use a wireless technology called IEEE 802.11 (a, b, g, etc.) to provide a secure, reliable, and high-speed wireless connection. Wi-Fi can be used to connect computers to each other, the Internet, and a wired network (using IEEE 802.3 or Ethernet). The Wi-Fi network may operate in unlicensed 2.4 and 5 GHz wireless bands, for example, at 11 Mbps (802.11a) or 54 Mbps (802.11b) data rates, or in products containing both bands (dual band) .

Those of ordinary skill in the art will understand that information and signals may be represented using any of a variety of different technologies and techniques. For example, data, instructions, commands, information, signals, bits, symbols, and chips that may be referenced in the above description may include voltages, currents, electromagnetic waves, magnetic fields or particles, Particles or particles, or any combination thereof.

Those skilled in the art will appreciate that the various illustrative logical blocks, modules, processors, means, circuits, and algorithm steps described in connection with the embodiments disclosed herein may be implemented or performed with a specific purpose, (Which may be referred to herein as "software") or a combination of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the design constraints imposed on the particular application and the overall system. Those skilled in the art may implement the described functions in various ways for each particular application, but such implementation decisions should not be interpreted as being outside the scope of the present disclosure.

The various embodiments presented herein may be implemented as a method, apparatus, or article of manufacture using standard programming and / or engineering techniques. The term "article of manufacture" includes a computer program, carrier, or media accessible from any computer-readable storage device. For example, the computer-readable storage medium can be a magnetic storage device (e.g., a hard disk, a floppy disk, a magnetic strip, etc.), an optical disk (e.g., CD, DVD, etc.) But are not limited to, memory devices (e. G., EEPROM, card, stick, key drive, etc.). The various storage media presented herein also include one or more devices and / or other machine-readable media for storing information.

It will be appreciated that the particular order or hierarchy of steps in the presented processes is an example of exemplary approaches. It will be appreciated that, based on design priorities, a particular order or hierarchy of steps in the processes may be rearranged within the scope of this disclosure. The appended method claims provide elements of the various steps in a sample order, but are not meant to be limited to the specific order or hierarchy presented.

The description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the scope of the present disclosure. Thus, the present disclosure should not be construed as limited to the embodiments set forth herein, but is to be accorded the widest scope consistent with the principles and novel features presented herein.

Claims (16)

21. A computer program stored in a computer-readable storage medium including encoded instructions for causing the one or more processors to perform the following steps for indexing, when executed by one or more processors of a computer system: Said steps comprising:
Receiving biometric data including one or more items and change data different from the biometric data;
Inputting the biometric data and change data to a trained neural network, and computing the biometric data and the change data using the trained neural network;
Generating first health risk information for the biometric data based on the output of the trained neural network and generating second health risk information for the change data; And
Generating a correlation between the data change amount of the biometric data and the change data, and the health risk information change amount of the first health risk information and the second health risk information;
/ RTI >
A computer program stored on a computer readable medium.
The method according to claim 1,
The biometric data includes:
At least one of demographic information, demographic information, death related information, health insurance type information, socioeconomic level information, disability registration information, physical examination information, history information, family history information, lifestyle information, Lt; / RTI >
A computer program stored on a computer readable storage medium.
The method according to claim 1,
Extracting a highly related item having a high correlation with the first health risk information among at least one item of the biometric data; And
Generating risk avoiding information for changing the highly correlated item;
≪ / RTI >
A computer program stored on a computer readable storage medium.
The method according to claim 1,
Wherein the first health risk information comprises:
A probability of onset of cardiovascular disease, an onset probability of cancer, and an onset probability of adult disease,
A computer program stored on a computer readable medium.
The method according to claim 1,
Determining threshold change data in which a ratio of the data change amount to the health risk information change amount is equal to or greater than a predetermined threshold value based on the association relationship;
≪ / RTI >
A computer program stored on a computer readable medium.
A computing device for predicting a health risk occurrence,
One or more processors; And
A memory for storing instructions executable on the one or more processors;
Lt; / RTI >
Wherein the one or more processors comprise:
Receiving biometric data including at least one item and change data in which at least one item of the biometric data is changed,
Inputting the biometric data and change data to a trained neural network, computing the biometric data and the change data using the trained neural network,
Generating first health risk information for the biometric data based on the output of the trained neural network, generating second health risk information for the change data, and
And generating a correlation between the data change amount of the biometrics data and the change data and the health risk information change amount of the first health risk information and the second health risk information,
A computing device for predicting health risks.
WHAT IS CLAIMED IS: 1. A method for predicting a health risk occurrence performed on one or more processors of a computing device,
Receiving biometric data including at least one item and change data in which at least one item of the biometric data is changed;
Inputting the biometric data and change data to a trained neural network, and computing the biometric data and the change data using the trained neural network;
Generating first health risk information for the biometric data based on the output of the trained neural network and generating second health risk information for the change data; And
Generating a correlation between the data change amount of the biometric data and the change data, and the health risk information change amount of the first health risk information and the second health risk information;
/ RTI >
A method for predicting a health risk occurrence performed on one or more processors of a computing device.
21. A computer program stored in a computer-readable storage medium including encoded instructions for causing the one or more processors to perform the following steps for indexing, when executed by one or more processors of a computer system: Said steps comprising:
Acquiring biometric data;
Grouping the biometrics data by objects based on object classification information included in the biometrics data to generate samples for each object;
Generating labeled learning data by matching the subject-specific sample with the clinical data; And
Generating a neural network model for training the neural network using the labeled learning data to output health risk information based on the biometric data;
/ RTI >
A computer program stored on a computer readable storage medium.
9. The method of claim 8,
The neural network is composed of an RNN (recurrent neural network)
Wherein the RNN comprises a long short-term memory (LSTM)
A computer program stored on a computer readable storage medium.
9. The method of claim 8,
Wherein the generating the neural network model comprises:
Inputting the sample to the neural network and comparing the output data calculated by the neural network with the medical care data to calculate an error;
Adjusting a weight of the neural network based on the error in a backpropagation manner;
Determining whether to stop learning using verification data if learning of the neural network is performed over a predetermined epoch; And
Testing the performance of the neural network model using test data and determining whether to activate the neural network based on the performance;
/ RTI >
A computer program stored on a computer readable storage medium.
9. The method of claim 8,
The labeling,
Including the occurrence of the disease included in the medical care data, the name of the disease,
The learning data
Which comprises many years of biometric data and which is classified as a negative sample or a positive sample depending on the occurrence of the disease,
A computer program stored on a computer readable storage medium.
12. The method of claim 11,
Dividing the biometric data of the multi-year period by the end point of the biometric data for a long period of time in the positive sample to generate an artificial positive sample; And
Dividing the multi-year biometric data of the speech sample so as to have an arbitrary start time and end time to generate artificial speech samples;
≪ / RTI >
A computer program stored on a computer readable storage medium.
9. The method of claim 8,
Normalizing a measurement value included in the biometric data and setting a missing value included in the biometric data as a total average value;
≪ / RTI >
A computer program stored on a computer readable storage medium.
9. The method of claim 8,
Wherein the acquiring of the biometrics data comprises:
Obtaining the biometric data from at least one of an electronic health record (EHR), an electronic medical record (EMR), and a health check database from a hospital server and a government server;
/ RTI >
A computer program stored on a computer readable storage medium.
A computing device for predicting a health risk occurrence,
One or more processors; And
A memory for storing instructions executable on the one or more processors;
Lt; / RTI >
Wherein the one or more processors comprise:
Acquiring biometric data,
Grouping the biometrics data by individuals based on the subject classification information included in the biometrics data to generate a subject-specific sample,
Generates learning data labeled by matching the subject-specific sample with the clinical data, and
And generating a neural network model for training the neural network using the labeled learning data to output health risk information based on the biometric data,
A computing device for predicting health risks.
WHAT IS CLAIMED IS: 1. A method for predicting a health risk occurrence performed on one or more processors of a computing device,
Acquiring biometric data;
Grouping the biometric data by individuals based on subject classification information included in the biometric data to generate a subject-specific sample;
Generating labeled learning data by matching the subject-specific sample with the clinical data; And
Generating a neural network model for training the neural network using the labeled learning data to output health risk information based on the biometric data;
/ RTI >
A method for predicting a health risk occurrence performed on one or more processors of a computing device.

Images