Medical indicator risk early warning method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a medical index risk early warning method and device, electronic equipment and a storage medium, and belongs to the technical field of medical science and technology and artificial intelligence. The method comprises the following steps: performing feature extraction on target medical index data to obtain medical index features; performing risk degree identification on the medical index features to obtain a medical index risk degree; calculating a target residual error between the medical index risk degree and the medical index observation degree through a residual error layer; and carrying out risk early warning on the target medical index data according to the target residual error to obtain a risk early warning level. According to the embodiment of the invention, by calculating the target residual error between the medical index risk degree and the observed medical index degree, the abnormal state of the medical index can be effectively identified, potential health problems can be found in time, risk early warning is carried out on the target medical index data according to the target residual error and the risk early warning layer, and the risk early warning efficiency is improved. Dynamic prediction of the medical indexes can be realized, and the accuracy of risk early warning of the medical indexes can be improved.

Description

Technical Field

[0001] The present application relates to the fields of medical technology and artificial intelligence technology, and in particular to a medical indicator risk warning method and device, electronic equipment and storage medium. Background Art

[0002] At present, the user's medical indicators are usually monitored through the periodic data collection method of smart watches, and the medical indicators are further analyzed and warned through the medical health detection system. For example, in the medical technology scenario, the user monitors his blood sugar and heart rate data through a smart watch. The smart watch automatically measures blood sugar every 2 hours and records heart rate data throughout the day. The built-in detection software of the smart watch uses a static risk threshold to assess the user's health status (such as fasting blood sugar above 7.0mmol / L is considered abnormal). However, this method relies on a single static risk threshold, and the collected medical indicators are not continuous data sets, which makes the risk analysis of the user's medical indicators relatively simple, resulting in a low accuracy rate of medical indicator risk warnings. Therefore, how to improve the accuracy of medical indicator risk warnings has become an urgent problem to be solved. Summary of the invention

[0003] The main purpose of the embodiments of the present application is to propose a medical indicator risk warning method and device, electronic equipment and storage medium, aiming to improve the accuracy of medical indicator risk warning.

[0004] To achieve the above-mentioned purpose, a first aspect of an embodiment of the present application proposes a medical indicator risk warning method, the method comprising:

[0005] Collect target medical indicator data of target users;

[0006] Obtaining a pre-trained risk warning model; wherein the risk warning model includes a coding layer, a decoding layer, a residual layer and a risk warning layer;

[0007] Extracting features of the target medical indicator data through the coding layer to obtain medical indicator features;

[0008] Identifying the risk level of the medical indicator feature through the decoding layer to obtain the medical indicator risk level of the target medical indicator data;

[0009] Obtaining the observed medical indicator level of the target medical indicator data, and calculating the target residual between the medical indicator risk level and the observed medical indicator level through the residual layer;

[0010] A risk warning is performed on the target medical indicator data according to the target residual and the risk warning layer to obtain a risk warning level.

[0011] In some embodiments, performing risk warning on the target medical indicator data according to the target residual and the risk warning layer to obtain a risk warning level includes:

[0012] Obtaining a risk warning threshold of the target medical indicator data;

[0013] The target medical indicator data is subjected to risk warning according to the risk warning threshold, the target residual and the risk warning layer to obtain the risk warning level.

[0014] In some embodiments, obtaining the risk warning threshold of the target medical indicator data includes:

[0015] Using a preset sliding window to obtain the window health and medical indicator data of the target user;

[0016] Performing indicator prediction on the window health and medical indicator data through the risk warning model to obtain window predicted health and medical indicator data;

[0017] Calculating the window health indicator residual between the window health indicator and the window predicted health indicator data, and combining the window health indicator residuals based on a preset window time sequence to obtain a window health indicator residual sequence;

[0018] The residual mean and residual standard deviation of the window health indicator residual sequence are obtained, and the risk warning threshold of the target medical indicator data is determined according to the multiples of the residual mean and the residual standard deviation.

[0019] In some embodiments, the step of identifying the risk level of the medical indicator feature through the decoding layer to obtain the medical indicator risk level of the target medical indicator data includes:

[0020] Mapping the medical indicator features to a preset high-dimensional space through the decoding layer to obtain high-dimensional indicator features;

[0021] The risk level of the high-dimensional indicator features is identified according to a preset reconstruction weight matrix to obtain the risk level of the medical indicator.

[0022] In some embodiments, extracting features of the target medical indicator data through the coding layer to obtain medical indicator features includes:

[0023] Extracting statistical distribution features of the target medical indicator data through the coding layer to obtain indicator statistical features;

[0024] Performing time series feature extraction on the indicator statistical features to obtain indicator time features;

[0025] Extracting frequency domain features of the indicator time features to obtain indicator frequency domain features;

[0026] The indicator statistical characteristics, the indicator time characteristics and the indicator frequency domain characteristics are integrated to obtain the medical indicator characteristics.

[0027] In some embodiments, collecting target medical indicator data of a target user includes:

[0028] Acquire the initial medical indicator data of the target user detected by the preset hardware device;

[0029] Performing abnormal value detection on the initial medical indicator data to obtain abnormal medical indicator data, and removing the abnormal medical indicator data from the standard medical indicator data to obtain normal medical indicator data;

[0030] Smoothing the normal medical indicator data to obtain smoothed medical indicator data;

[0031] The smoothed medical indicator data is subjected to data filling processing to obtain the target medical indicator data.

[0032] In some embodiments, after performing risk warning on the target medical indicator data according to the target residual and the risk warning layer to obtain the risk warning level, the method further includes:

[0033] Obtaining the residual sum of squares of each target residual;

[0034] Obtaining the risk contribution of each target medical indicator data according to each target residual and the residual square sum;

[0035] The target risk contribution with the largest contribution is screened out from the risk contributions, and the target medical indicator data of the target risk contribution is used as the risk cause for the risk warning level.

[0036] To achieve the above-mentioned purpose, a second aspect of an embodiment of the present application proposes a medical indicator risk warning device, the device comprising:

[0037] A medical indicator acquisition module is used to collect target medical indicator data of target users;

[0038] A model acquisition module is used to acquire a pre-trained risk warning model; wherein the risk warning model includes a coding layer, a decoding layer, a residual layer and a risk warning layer;

[0039] A feature extraction module, used for extracting features of the target medical indicator data through the coding layer to obtain medical indicator features;

[0040] A risk level identification module, used to identify the risk level of the medical indicator feature through the decoding layer to obtain the medical indicator risk level of the target medical indicator data;

[0041] A residual calculation module, used for obtaining the observed medical indicator level of the target medical indicator data, and calculating the target residual between the medical indicator risk level and the observed medical indicator level through the residual layer;

[0042] The risk warning module is used to perform risk warning on the target medical indicator data according to the target residual and the risk warning layer to obtain the risk warning level.

[0043] To achieve the above objectives, a third aspect of an embodiment of the present application proposes an electronic device, which includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method described in the first aspect is implemented.

[0044] To achieve the above-mentioned purpose, the fourth aspect of an embodiment of the present application proposes a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0045] The medical indicator risk warning method and device, electronic device and storage medium proposed in the present application collect target medical indicator data of target users; obtain a pre-trained risk warning model; wherein the risk warning model includes a coding layer, a decoding layer, a residual layer and a risk warning layer; extract features of the target medical indicator data through the coding layer to obtain medical indicator features; identify the risk level of the medical indicator features through the decoding layer to obtain the medical indicator risk level of the target medical indicator data; obtain the observed medical indicator level of the target medical indicator data, and calculate the target residual between the medical indicator risk level and the observed medical indicator level through the residual layer; perform risk warning on the target medical indicator data according to the target residual and the risk warning layer to obtain a risk warning level. This application first extracts features from target medical indicator data through a coding layer to obtain medical indicator features, and identifies the risk levels of the medical indicator features through a decoding layer, thereby being able to identify and capture key information about medical indicators, and predict individual health risks based on their indicator features, thereby achieving personalized health monitoring; secondly, by calculating the target residual between the risk level of medical indicators and the observed medical indicator level through a residual layer, the abnormal state of medical indicators can be effectively identified, and potential health problems can be discovered in a timely manner; finally, risk warnings are issued for target medical indicator data based on the target residual and risk warning layer, thereby achieving dynamic prediction of medical indicators, and helping to improve the accuracy of risk warnings for medical indicators. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 is a flow chart of a medical indicator risk warning method provided in an embodiment of the present application;

[0047] Figure 2 yes Figure 1 Flow chart of step S101 in FIG.

[0048] Figure 3 yes Figure 1 Flow chart of step S103 in FIG.

[0049] Figure 4 yes Figure 1 Flow chart of step S104 in FIG.

[0050] Figure 5 yes Figure 1 Flow chart of step S106 in FIG.

[0051] Figure 6 yes Figure 5 Flow chart of step S501 in FIG.

[0052] Figure 7 is another flow chart of the medical indicator risk warning method provided in an embodiment of the present application;

[0053] Figure 8 It is a structural schematic diagram of a medical indicator risk warning device provided in an embodiment of the present application;

[0054] Fig. 9 It is a schematic diagram of the hardware structure of the electronic device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0056] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0057] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0058] First, some nouns involved in this application are analyzed:

[0059] Artificial intelligence (AI) is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. AI is a branch of computer science that attempts to understand the essence of intelligence and produce a new type of intelligent machine that can respond in a similar way to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing and expert systems. AI can simulate the information process of human consciousness and thinking. AI is also a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0060] Based on this, the embodiments of the present application provide a medical indicator risk warning method and device, an electronic device and a storage medium, aiming to improve the accuracy of medical indicator risk warning.

[0061] The medical indicator risk warning method and device, electronic device and storage medium provided in the embodiments of the present application are specifically illustrated through the following embodiments. First, the medical indicator risk warning method in the embodiments of the present application is described.

[0062] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0063] The basic technologies of artificial intelligence generally include sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, processing / interaction systems, mechatronics, etc. Artificial intelligence software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0064] The medical indicator risk warning method provided in the embodiment of the present application relates to the fields of medical technology and artificial intelligence technology. The medical indicator risk warning method provided in the embodiment of the present application can be applied to the terminal, can also be applied to the server side, and can also be software running in the terminal or the server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or it can be configured as a server cluster or distributed system composed of multiple physical servers, and can also be configured as a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms and other basic cloud computing services; the software can be an application that implements the medical indicator risk warning method, etc., but is not limited to the above forms.

[0065] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0066] It should be noted that in each specific implementation of the present application, when it comes to the need to perform relevant processing based on the target medical indicator data of the target user and other data related to the user's identity or characteristics, the user's permission or consent will be obtained first, and the collection, use and processing of these data will comply with relevant laws, regulations and standards. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or consent will be obtained through a pop-up window or jump to a confirmation page. After clearly obtaining the user's separate permission or consent, the necessary user-related data for the normal operation of the embodiment of the present application will be obtained.

[0067] Figure 1 is an optional flow chart of the medical indicator risk warning method provided in the embodiment of the present application. Figure 1 The method may include but is not limited to steps S101 to S106.

[0068] Step S101, collecting target medical indicator data of target users.

[0069] Step S102, obtaining a pre-trained risk warning model; wherein the risk warning model includes a coding layer, a decoding layer, a residual layer and a risk warning layer.

[0070] Step S103, extracting features of the target medical indicator data through the coding layer to obtain medical indicator features.

[0071] Step S104, identifying the risk level of the medical indicator features through the decoding layer to obtain the medical indicator risk level of the target medical indicator data.

[0072] Step S105, obtaining the observed medical indicator level of the target medical indicator data, and calculating the target residual between the medical indicator risk level and the observed medical indicator level through the residual layer.

[0073] Step S106, performing risk warning on the target medical indicator data according to the target residual and the risk warning layer to obtain the risk warning level.

[0074] Steps S101 to S106 shown in the embodiment of the present application are performed by collecting target medical indicator data of the target user; obtaining a pre-trained risk warning model; wherein the risk warning model includes a coding layer, a decoding layer, a residual layer and a risk warning layer; performing feature extraction on the target medical indicator data through the coding layer to obtain medical indicator features; performing risk level identification on the medical indicator features through the decoding layer to obtain the medical indicator risk level of the target medical indicator data; obtaining the observed medical indicator level of the target medical indicator data, and calculating the target residual between the medical indicator risk level and the observed medical indicator level through the residual layer; performing risk warning on the target medical indicator data according to the target residual and the risk warning layer to obtain a risk warning level. This application first extracts features from target medical indicator data through a coding layer to obtain medical indicator features, and identifies the risk levels of the medical indicator features through a decoding layer, thereby being able to identify and capture key information about medical indicators, and predict individual health risks based on their indicator features, thereby achieving personalized health monitoring; secondly, by calculating the target residual between the risk level of medical indicators and the observed medical indicator level through a residual layer, the abnormal state of medical indicators can be effectively identified, and potential health problems can be discovered in a timely manner; finally, risk warnings are issued for target medical indicator data based on the target residual and risk warning layer, thereby achieving dynamic prediction of medical indicators, and helping to improve the accuracy of risk warnings for medical indicators.

[0075] In step S101 of some embodiments, specifically, the target medical indicator data refers to physiological parameters of the target user.

[0076] For example, parameters such as heart rate, blood pressure, blood oxygen saturation, body temperature, sleep conditions, stress and blood sugar.

[0077] See also Figure 2 In some embodiments, step S101 includes but is not limited to steps S201 to S204:

[0078] Step S201, obtaining the initial medical indicator data of the target user detected by the preset hardware device.

[0079] Step S202, performing abnormal value detection on the initial medical indicator data to obtain abnormal medical indicator data, and removing the abnormal medical indicator data from the standard medical indicator data to obtain normal medical indicator data.

[0080] Step S203, smoothing the normal medical indicator data to obtain smoothed medical indicator data.

[0081] Step S204, performing data filling processing on the smoothed medical indicator data to obtain target medical indicator data.

[0082] In step S201 of some embodiments, the collection of initial medical indicators can be implemented by smart hardware devices, wherein the smart hardware devices can be wearable devices (such as smart watches, smart bracelets) or portable monitoring instruments (such as blood pressure monitors, blood glucose meters), etc.

[0083] Specifically, the smart hardware device can continuously collect the physiological parameters of the target user within 24 hours and has a certain local storage capacity so that it can continue to work even when the network is unstable.

[0084] Specifically, various sensors are integrated into smart hardware devices, such as photoelectric heart rate sensors, blood pressure sensors, and blood oxygen sensors. Among them, the light beam emitted by the photoelectric heart rate sensor illuminates the skin, and the light absorption changes caused by blood flow can be measured to monitor heart rate and heart rate variability; the pressure changes caused by blood reflow can be detected by the blood pressure sensor to calculate systolic and diastolic pressures; the blood oxygen sensor can emit two different wavelengths of light beams, which penetrate the skin and blood vessels and are absorbed by hemoglobin. The absorption of different wavelengths of light by hemoglobin is measured to determine the user's blood oxygen saturation.

[0085] For example, in a medical technology scenario, the target user wears a smart watch that integrates the above-mentioned sensors. At different times of the day, the smart watch will automatically wake up the sensor, collect initial medical indicator data at preset time intervals (such as once a minute), and process these initial medical indicator data through the microprocessor in the smart watch to calculate the user's heart rate, blood pressure, blood oxygen saturation, body temperature and other physiological parameters.

[0086] In this embodiment, by using high-precision sensors such as photoelectric heart rate sensors, blood pressure sensors, and blood oxygen sensors to collect target medical indicator data, the accuracy of the collected target medical indicator data can be ensured.

[0087] In an optional embodiment of the present application, the sensor data can also be calibrated and compensated through the built-in algorithm of the intelligent hardware device to eliminate environmental interference and equipment errors, and sensors such as accelerometers and gyroscopes can be used to further monitor the motion state of the target user, and the collection strategy of the target medical indicator data can be adjusted according to the motion state to improve the data quality in the motion state.

[0088] In step S202 of some embodiments, specifically, outlier detection may be performed on the initial medical indicator data using statistical methods (such as standard deviation, box plot).

[0089] Specifically, first obtain the target user's average medical indicator and standard deviation of the medical indicator within one day, and compare the average medical indicator with the standard deviation of the medical indicator. If the initial medical indicator exceeds the average medical indicator and is much smaller than the standard deviation of the medical indicator or much larger than the standard deviation of the medical indicator, the initial medical indicator is determined to be an abnormal indicator.

[0090] For example, in the process of collecting heart rate medical indicators, if a heart rate data point exceeds 3 standard deviations of the average heart rate value, we can mark the heart rate data point as an outlier and remove it from the collected heart rate medical indicator data.

[0091] In step S203 of some embodiments, specifically, the normal medical indicator data may be smoothed by a moving average or exponential smoothing method.

[0092] Specifically, based on the preset smoothing window size, for each normal medical indicator data collected, the average medical indicator value covered by the previous and next smoothing window sizes is calculated, and the medical average value is used as the smoothed medical indicator data.

[0093] For example, if the smoothing window size is 5, the heart rate smoothing value is calculated every 5 seconds, and the heart rates of 72bpm, 70bpm, 68bpm, 76bpm, and 74bpm collected within 0-5s are averaged to obtain a heart rate of 72.4bpm as the smoothed medical indicator data.

[0094] In this embodiment, by smoothing normal medical indicator data to obtain smoothed medical indicator data, random fluctuations or noise in the collected medical indicator data can be reduced, thereby making the trends and patterns of the medical indicator data more obvious and improving the data quality of the collected medical indicator data.

[0095] In step S204 of some embodiments, specifically, filling can be performed by linear interpolation, moving average or time series interpolation method.

[0096] For example, for the collected heart rate monitoring data, the heart rate data is recorded every 30 minutes in a day, but due to sensor failure, some data is missing. The heart rate collected at 0.5h is 68bpm, the heart rate data collected at 1h is indeed, and the heart rate collected at 1.5h is 77bpm. The heart rate for 1h is calculated by linear interpolation to be 71.6bpm.

[0097] In an optional embodiment of the present application, the target medical indicator data can also be stored in the memory of the smart hardware device, and the target medical indicator data can be transmitted to the target user's mobile terminal or cloud server via wireless communication technologies such as Bluetooth or Wi-Fi.

[0098] In step S102 of some embodiments, specifically, the risk warning model is a neural network model composed of a coding layer, a decoding layer, a residual layer and a risk warning layer.

[0099] Furthermore, the encoding layer is used to extract the characteristics of the target medical indicator data; the decoding layer is used to identify the risk level of the medical indicator characteristics, predict health risks based on individual indicator characteristics, and realize personalized health monitoring; the residual layer is used to calculate the target residual between the risk level of the medical indicator and the observed medical indicator level, and effectively identify the abnormal state of the medical indicator; the risk warning layer is used to issue risk warnings for the target medical indicator data, and realize dynamic prediction of medical indicators.

[0100] See also Figure 3 In some embodiments, step S103 includes but is not limited to steps S301 to S304:

[0101] Step S301, extracting statistical distribution features of target medical indicator data through a coding layer to obtain indicator statistical features.

[0102] Step S302: extract time series features from the indicator statistical features to obtain indicator time features.

[0103] Step S303: extract frequency domain features from the indicator time features to obtain indicator frequency domain features.

[0104] Step S304, integrating the indicator statistical characteristics, indicator time characteristics and indicator frequency domain characteristics to obtain the medical indicator characteristics.

[0105] In step S301 of some embodiments, specifically, the indicator statistical characteristics refer to statistical data characteristics of the target medical indicator data, and the indicator statistical characteristics include but are not limited to mean, median, standard deviation, skewness and kurtosis characteristics, etc.

[0106] Specifically, the encoding layer may include multiple hidden layers, and by adjusting the weights and bias parameters of the hidden layers, features of different dimensions may be extracted.

[0107] Specifically, the statistical distribution characteristics of the target medical indicator data can be extracted by the following formula:

[0108] h1=σ(W1X+b1)

[0109] Among them, h1 represents the statistical characteristics of the indicator, σ represents the activation function, W1 represents the statistical weight, X represents the target medical indicator data, and b1 represents the statistical bias parameter.

[0110] In step S302 of some embodiments, specifically, the indicator time feature refers to a feature extracted from the time series data of the target medical indicator data, such as autocorrelation, periodic components, trends, etc.

[0111] Specifically, the frequency domain feature extraction of the indicator time feature can be performed using the following formula:

[0112] h2=σ(W2h1+b2)

[0113] Among them, h2 represents the time characteristic of the indicator, σ represents the activation function, W2 represents the time weight, h1 represents the statistical characteristic of the indicator, and b2 represents the time bias parameter.

[0114] In step S303 of some embodiments, specifically, the index frequency domain feature refers to extracting features such as amplitude and phase of target medical index data.

[0115] Specifically, the time series data is converted to the frequency domain by fast Fourier transform before extracting the frequency domain features.

[0116] Specifically, the frequency domain feature extraction of the indicator time feature can be performed using the following formula:

[0117] h3=σ(W3h2+b3)

[0118] Among them, h3 represents the frequency domain characteristics of the indicator, σ represents the activation function, W3 represents the frequency domain weight, h2 represents the time characteristics of the indicator, and b3 represents the frequency domain bias parameter.

[0119] In step S304 of some embodiments, specifically, the indicator statistical features, indicator time features and indicator frequency domain features are spliced, and the spliced ​​features are activated by a preset activation function (such as a Sigmoid function) to obtain the medical indicator features represented in a low-dimensional space.

[0120] For example, in the medical technology scenario, if the target medical indicator data is heart rate monitoring data collected in real time, the encoding layer first learns the statistical indicator features of the daily mean heart rate and the heart rate standard deviation; secondly, the daily periodicity of the heart rate can be identified by extracting the time indicator features, and the main amplitude and phase components of the heart rate changes can be further extracted to determine the frequency of the heart rate changes; finally, these features are fused to form a medical indicator feature that contains all features.

[0121] In this embodiment, by fusing the indicator statistical characteristics, indicator time characteristics and indicator frequency domain characteristics, multi-level indicator characteristics can be extracted from the target medical indicator data to accurately identify and capture the main key information of the medical indicator, thereby improving the accuracy of medical indicator feature extraction and facilitating the subsequent improvement of the accuracy of medical indicator risk warning.

[0122] See also Figure 4 In some embodiments, step S104 includes but is not limited to steps S401 to S402:

[0123] Step S401, mapping the medical indicator features to a preset high-dimensional space through a decoding layer to obtain high-dimensional indicator features.

[0124] Step S402, identifying the risk level of the high-dimensional indicator features according to the preset reconstruction weight matrix to obtain the risk level of the medical indicator.

[0125] In step S401 of some embodiments, specifically, the high-dimensional indicator feature refers to the feature of the medical indicator feature mapped to the high-dimensional space.

[0126] Specifically, decoding weights and decoding bias parameters are obtained. Based on the decoding weights and decoding bias parameters, the medical indicator features output by the encoding layer can be mapped to a high-dimensional space, and the interaction between the medical indicator features can be further learned through nonlinear activation functions (such as ReLU, sigmoid or tanh functions).

[0127] For example, there is an interactive relationship between heart rate and blood oxygen saturation, and a decrease in blood oxygen saturation may cause an increase in heart rate; there is an interactive relationship between blood pressure and heart rate, and heart rate will increase during strenuous exercise or stress changes.

[0128] In step S402 of some embodiments, specifically, a reconstructed weight matrix is ​​obtained by performing nonlinear activation on the real health indicator matrix and the observed medical indicator matrix of the target user, wherein the real health indicator matrix is ​​used to describe the health and medical indicator values ​​of the target user in the historical time period; and the observed medical indicator matrix is ​​used to describe the medical indicator values ​​of the target user in the observed time period.

[0129] Specifically, by reconstructing the weight matrix to identify the risk level of high-dimensional indicator features, the preliminary medical indicator risk level of the target user can be preliminarily determined.

[0130] For example, in medical technology scenarios, heart rate and blood oxygen saturation are normal. A decrease in blood oxygen saturation may cause an increase in heart rate as a compensation mechanism for the body to improve oxygen delivery. If the blood oxygen saturation reading is continuously below 90% during sleep without a corresponding increase in heart rate, it indicates that the target user's heart rate or blood oxygen saturation indicators are at risk.

[0131] In step S105 of some embodiments, specifically, the target residual is used to describe the difference between the medical indicator risk level and the observed medical indicator level, and the abnormal state of the medical indicator of the target user can be determined through the target residual.

[0132] Specifically, the observed medical indicator level refers to the health level of the target medical indicator data collected in real time.

[0133] Specifically, if the medical indicator risk level is combined with the real medical indicator health level in the historical period, and is close to the observed medical indicator level, it means that the medical indicator risk level prediction accuracy is high. However, when the medical indicator has a greater risk, the medical indicator risk level is combined with the real medical indicator health level, and is significantly different from the observed medical indicator level, indicating that the medical indicator risk level prediction accuracy is low. Based on this, the size of the target residual can reflect whether the target medical indicator data has a health risk.

[0134] See also Figure 5 In some embodiments, step S106 includes but is not limited to steps S501 to S502:

[0135] Step S501, obtaining the risk warning threshold of the target medical indicator data.

[0136] Step S502, performing risk warning on the target medical indicator data according to the risk warning threshold, target residual and risk warning layer to obtain the risk warning level.

[0137] See also Figure 6 In some embodiments, step S501 includes but is not limited to steps S601 to S604:

[0138] Step S601, using a preset sliding window to obtain window health and medical indicator data of a target user.

[0139] Step S602: predicting the window health and medical indicator data through the risk warning model to obtain the window predicted health and medical indicator data.

[0140] Step S603, calculating the window health indicator residual between the window health indicator and the window predicted health indicator data, and combining the window health indicator residuals based on a preset window time sequence to obtain a window health indicator residual sequence.

[0141] Step S604, obtaining the residual mean and residual standard deviation of the window health indicator residual sequence, and determining the risk warning threshold of the target medical indicator data according to the multiples of the residual mean and the residual standard deviation.

[0142] In step S601 of some embodiments, specifically, the preset sliding window is based on the collection frequency and expected time range of the target medical indicator data.

[0143] For example, the target user's healthy heart rate indicator collected every ten minutes is used as the sliding window size, and the healthy heart rate collected once a minute is the step size of the sliding window movement, and the healthy heart rate indicator collected every ten minutes is used as the window health medical indicator data.

[0144] In step S602 of some embodiments, specifically, the health level of the window health and medical indicator data is predicted by the risk warning model to obtain the window predicted health and medical indicator data.

[0145] For example, the heart rate and blood pressure data in each 10-minute window are input into the risk warning model, and the model outputs the corresponding predicted heart rate and blood pressure data.

[0146] In step S603 of some embodiments, specifically, for each sliding window, the actual window health and medical indicator data is subtracted from the window predicted health and medical indicator data to obtain the window health indicator residual corresponding to each sliding window, which is used to describe the risk situation of the health and medical indicator data collected by each sliding window.

[0147] Specifically, according to the time sequence of the windows, the window health indicator residuals of all windows are combined to obtain a window health indicator residual sequence.

[0148] In step S604 of some embodiments, specifically, the statistics of the window health indicator residual sequence, such as the mean and standard deviation, are calculated, and the risk warning threshold is determined based on the multiples of the mean and standard deviation; wherein, the multiples of the residual standard deviation can be determined based on the actual application scenario, and by selecting different multiples, it is possible to quantify different degrees of medical indicator abnormalities.

[0149] For example, the risk warning threshold can be a mild risk warning threshold, a moderate risk warning threshold and a severe risk warning threshold. Since the distribution of the residual sequence of the window health indicator is approximately normal, in order to simplify the calculation process, the residual mean can be determined to be 0 and the residual standard deviation to be 1. The mild risk warning threshold is the residual mean plus 1.5 times the residual standard deviation: 0+1x1.5=1.5; the moderate risk warning threshold is the residual mean plus 2 times the residual standard deviation: 0+1x2=2; the severe risk warning threshold is the residual mean plus 2.5 times the residual standard deviation: 0+1x2.5=2.5.

[0150] In this embodiment, the risk warning threshold of the target medical indicator data is determined based on the multiples of the residual mean and the residual standard deviation. The risk warning threshold can be dynamically adjusted according to the changing trend of the medical indicator data of the target user to adapt to the changes in medical indicators at any time. The corresponding risk warning threshold can be matched for different users, providing users with personalized risk predictions, which will help to subsequently improve the accuracy of medical indicator risk warnings.

[0151] In step S502 of some embodiments, specifically, the risk warning level is a warning level divided based on the risk warning threshold and the target residual.

[0152] Specifically, the risk warning level may also include a mild risk warning level, a moderate risk warning level, and a severe risk warning level. Among them, for the mild risk warning level, the target user's medical indicators can be reminded that there are abnormalities, and the target user can adjust the medical indicators back to normal by improving diet or work and rest; for the moderate risk warning level, the target user can be reminded that the abnormality of the medical indicators is more serious, and further auxiliary pathological examination is required in the hospital; for the severe risk warning level, the target user can be reminded that the abnormality of the indicators may be a serious disease, and the disease diagnosis needs to be made in the hospital as soon as possible, and an emergency notification will be automatically sent to the user's pre-set emergency contact, providing the user's location information for emergency rescue, and providing the user with basic emergency handling guidance to help the user or the emergency contact take correct first aid measures before the arrival of professional medical personnel.

[0153] For example, if the residual of the monitored heart rate indicator does not exceed the risk warning threshold of 1.5, it means that all medical indicators of the target user are healthy; if the residual of the heart rate indicator exceeds the risk warning threshold of 1.5 but is less than 2, it means that the medical indicators of the target user belong to the mild risk warning level; if the residual of the heart rate indicator exceeds the risk warning threshold of 2 but is less than 2.5, it means that the medical indicators of the target user belong to the moderate risk warning level; if the residual of the heart rate indicator exceeds the risk warning threshold of 2.5, it means that the medical indicators of the target user belong to the severe risk warning level.

[0154] In this embodiment, risk warning is performed on the target medical indicator data according to the risk warning threshold, target residual and risk warning layer to obtain the risk warning level. The risk warning threshold and target residual can be combined with dynamic adjustment to predict the risk of medical indicators, thereby realizing dynamic prediction of medical indicators. Through multi-level risk classification, the risk degree of medical indicators is accurately quantified, which significantly improves the accuracy of medical indicator risk warning.

[0155] In an optional embodiment of the present application, the medical indicator data and trends of the target user can also be intuitively displayed in the user interface in the form of charts, etc.

[0156] Furthermore, when displaying medical indicators in the user interface, determine the medical indicator data that needs to be displayed (such as heart rate, blood pressure, blood sugar, etc.), and select the appropriate chart type according to the medical indicator data and the purpose of display (such as time series data is suitable for line charts, distribution data is suitable for bar charts or pie charts), and convert the medical indicator data into charts through data visualization tools or libraries (such as D3.js, Chart.js, Matplotlib, etc.). In the user interface, target users can also create charts according to their own needs.

[0157] For example, the mobile health application for diabetic patients provides a concise dashboard where patients can view health data including current blood sugar levels, activity levels and medication reminders, and view more detailed data by clicking different icons or buttons. In order to improve ease of use, the user interface also provides patients with large fonts, clear icons and a simple navigation structure to ensure that patients can easily use the application even if they have poor eyesight or are unfamiliar with the technology.

[0158] See also Figure 7 In some embodiments, after step S106, the medical indicator risk warning method further includes but is not limited to steps S701 to 703:

[0159] Step S701, obtaining the residual sum of squares of each target residual.

[0160] Step S702, obtaining the risk contribution of each target medical indicator data according to each target residual and the residual square sum.

[0161] Step S703, screening out the target risk contribution with the largest contribution from the risk contribution, and using the target medical indicator data of the target risk contribution as the risk cause for the risk warning level.

[0162] In step S701 of some embodiments, specifically, the residual sum of squares refers to the Euclidean norm representation of the target medical indicator for each target user.

[0163] For example, in the medical technology scenario, if the observed medical indicators of the target user are monitored as [72, 130 / 85, 5.5] (heart rate 72bpm, blood pressure 130 / 85mmHg, blood sugar 5.5mmol / L), and the predicted medical system is [70, 120 / 80, 5.0], then the target residual is [2, 10 / 5, 0.5]. The square of 2, 10 / 5 and 0.5 are summed, and the square root of the sum is taken to obtain the residual sum of squares.

[0164] In step S702 of some embodiments, specifically, the risk contribution degree refers to the ratio of the target residual of the target medical indicator to the residual sum of squares, which is used to indicate the contribution degree of each target medical indicator to the residual size.

[0165] For example, the ratio of the square of the risk contribution of heart rate is 2 to the residual sum of squares is 0.2; the ratio of the square of the risk contribution of blood pressure is 0 / 5 to the residual sum of squares is 0.6; the ratio of the square of the risk contribution of blood sugar is 0.5 to the residual sum of squares is 0.2.

[0166] In step S703 of some embodiments, for example, the risk contribution of heart rate is 0.2, the risk contribution of blood pressure is 0.6, and the risk contribution of blood sugar is 0.2, then the risk contribution of blood pressure is used as the target risk contribution, and the blood pressure index is used as the risk cause for the risk warning level.

[0167] In this embodiment, by screening out the target risk contribution with the largest contribution from the risk contribution, and using the target medical indicator data of the target risk contribution as the risk cause for the risk warning level, by combining the residual and the risk contribution, not only can a specific risk warning be generated, but also the potential causes of the risk warning can be pointed out, thereby providing more targeted health advice for the target users.

[0168] The embodiment of the present application collects target medical indicator data of the target user; obtains a pre-trained risk warning model; wherein the risk warning model includes a coding layer, a decoding layer, a residual layer and a risk warning layer; performs feature extraction on the target medical indicator data through the coding layer to obtain medical indicator features; performs risk level identification on the medical indicator features through the decoding layer to obtain the medical indicator risk level of the target medical indicator data; obtains the observed medical indicator level of the target medical indicator data, and calculates the target residual between the medical indicator risk level and the observed medical indicator level through the residual layer; performs risk warning on the target medical indicator data according to the target residual and the risk warning layer to obtain a risk warning level. This application first extracts features from target medical indicator data through a coding layer to obtain medical indicator features, and identifies the risk levels of the medical indicator features through a decoding layer, thereby being able to identify and capture key information about medical indicators, and predict individual health risks based on their indicator features, thereby achieving personalized health monitoring; secondly, by calculating the target residual between the risk level of medical indicators and the observed medical indicator level through a residual layer, the abnormal state of medical indicators can be effectively identified, and potential health problems can be discovered in a timely manner; finally, risk warnings are issued for target medical indicator data based on the target residual and risk warning layer, thereby achieving dynamic prediction of medical indicators, and helping to improve the accuracy of risk warnings for medical indicators.

[0169] See also Figure 8 The embodiment of the present application also provides a medical indicator risk warning device, which can implement the above-mentioned medical indicator risk warning method, and the device includes:

[0170] A medical indicator acquisition module is used to collect target medical indicator data of target users;

[0171] A model acquisition module is used to acquire a pre-trained risk warning model; wherein the risk warning model includes a coding layer, a decoding layer, a residual layer and a risk warning layer;

[0172] A feature extraction module is used to extract features of target medical indicator data through a coding layer to obtain medical indicator features;

[0173] A risk level identification module is used to identify the risk level of medical indicator features through a decoding layer to obtain the medical indicator risk level of target medical indicator data;

[0174] A residual calculation module, used to obtain the observed medical indicator level of the target medical indicator data, and calculate the target residual between the medical indicator risk level and the observed medical indicator level through the residual layer;

[0175] The risk warning module is used to conduct risk warning on the target medical indicator data according to the target residual and the risk warning layer to obtain the risk warning level.

[0176] The specific implementation of the medical indicator risk warning device is basically the same as the specific implementation of the above-mentioned medical indicator risk warning method, and will not be repeated here.

[0177] The embodiment of the present application also provides an electronic device, the electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned medical indicator risk warning method when executing the computer program. The electronic device can be any smart terminal including a tablet computer, a car computer, etc.

[0178] See also Fig. 9 , Fig. 9 The hardware structure of an electronic device of another embodiment is illustrated, and the electronic device includes:

[0179] The processor 901 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0180] The memory 902 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 902 can store processing systems and other applications. When the technical solution provided in the embodiments of this specification is implemented by software or firmware, the relevant program code is stored in the memory 902, and the processor 901 calls and executes the medical indicator risk warning method of the embodiment of this application;

[0181] Input / output interface 903, used to implement information input and output;

[0182] Communication interface 904, used to realize communication interaction between the device and other devices, which can be realized by wired mode (such as USB, network cable, etc.) or wireless mode (such as mobile network, WIFI, Bluetooth, etc.);

[0183] A bus 905 that transmits information between various components of the device (e.g., the processor 901, the memory 902, the input / output interface 903, and the communication interface 904);

[0184] The processor 901 , the memory 902 , the input / output interface 903 and the communication interface 904 are connected to each other in communication within the device via a bus 905 .

[0185] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned medical indicator risk warning method is implemented.

[0186] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0187] The medical indicator risk warning method, medical indicator risk warning device, electronic device and storage medium provided in the embodiments of the present application collect target medical indicator data of target users; obtain a pre-trained risk warning model; wherein the risk warning model includes a coding layer, a decoding layer, a residual layer and a risk warning layer; extract features of the target medical indicator data through the coding layer to obtain medical indicator features; identify the risk level of the medical indicator features through the decoding layer to obtain the medical indicator risk level of the target medical indicator data; obtain the observed medical indicator level of the target medical indicator data, and calculate the target residual between the medical indicator risk level and the observed medical indicator level through the residual layer; perform risk warning on the target medical indicator data according to the target residual and the risk warning layer to obtain a risk warning level. This application first extracts features from target medical indicator data through a coding layer to obtain medical indicator features, and identifies the risk levels of the medical indicator features through a decoding layer, thereby being able to identify and capture key information about medical indicators, and predict individual health risks based on their indicator features, thereby achieving personalized health monitoring; secondly, by calculating the target residual between the risk level of medical indicators and the observed medical indicator level through a residual layer, the abnormal state of medical indicators can be effectively identified, and potential health problems can be discovered in a timely manner; finally, risk warnings are issued for target medical indicator data based on the target residual and risk warning layer, thereby achieving dynamic prediction of medical indicators, and helping to improve the accuracy of risk warnings for medical indicators.

[0188] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0189] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0190] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0191] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0192] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0193] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0194] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. The mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0195] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0196] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0197] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0198] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention should be within the scope of the rights of the present invention.

Claims

1. A medical indicator risk early warning method, characterized in that: The method comprises: Collect target medical indicator data of target users; Obtaining a pre-trained risk warning model; wherein the risk warning model includes a coding layer, a decoding layer, a residual layer and a risk warning layer; Extracting features of the target medical indicator data through the coding layer to obtain medical indicator features; Identifying the risk level of the medical indicator feature through the decoding layer to obtain the medical indicator risk level of the target medical indicator data; Obtaining the observed medical indicator level of the target medical indicator data, and calculating the target residual between the medical indicator risk level and the observed medical indicator level through the residual layer; The target medical indicator data is subjected to risk warning according to the target residual and the risk warning layer to obtain a risk warning level.

2. The method according to claim 1, characterized in that The step of performing risk warning on the target medical indicator data according to the target residual and the risk warning layer to obtain a risk warning level includes: Obtaining a risk warning threshold of the target medical indicator data; The target medical indicator data is subjected to risk warning according to the risk warning threshold, the target residual and the risk warning layer to obtain the risk warning level.

3. The method according to claim 2, characterized in that The step of obtaining the risk warning threshold of the target medical indicator data includes: Using a preset sliding window to obtain the window health and medical indicator data of the target user; Performing indicator prediction on the window health and medical indicator data through the risk warning model to obtain window predicted health and medical indicator data; Calculating the window health indicator residual between the window health indicator and the window predicted health indicator data, and combining the window health indicator residuals based on a preset window time sequence to obtain a window health indicator residual sequence; The residual mean and residual standard deviation of the window health indicator residual sequence are obtained, and the risk warning threshold of the target medical indicator data is determined according to the multiples of the residual mean and the residual standard deviation.

4. The method according to claim 1, characterized in that: The step of identifying the risk level of the medical indicator feature through the decoding layer to obtain the medical indicator risk level of the target medical indicator data includes: Mapping the medical indicator features to a preset high-dimensional space through the decoding layer to obtain high-dimensional indicator features; The risk level of the high-dimensional indicator features is identified according to a preset reconstruction weight matrix to obtain the risk level of the medical indicator.

5. The method according to claim 1, characterized in that The step of extracting features of the target medical indicator data through the coding layer to obtain medical indicator features includes: Extracting statistical distribution features of the target medical indicator data through the coding layer to obtain indicator statistical features; Performing time series feature extraction on the indicator statistical features to obtain indicator time features; Extracting frequency domain features of the indicator time features to obtain frequency domain features of the indicator; The indicator statistical characteristics, the indicator time characteristics and the indicator frequency domain characteristics are integrated to obtain the medical indicator characteristics.

6. The method according to claim 1, characterized in that The collecting of target medical indicator data of the target user includes: Acquire the initial medical indicator data of the target user detected by the preset hardware device; Performing abnormal value detection on the initial medical indicator data to obtain abnormal medical indicator data, and removing the abnormal medical indicator data from the standard medical indicator data to obtain normal medical indicator data; Smoothing the normal medical indicator data to obtain smoothed medical indicator data; The smoothed medical indicator data is subjected to data filling processing to obtain the target medical indicator data.

7. The method according to any one of claims 1 to 6, characterized in that: After performing risk warning on the target medical indicator data according to the target residual and the risk warning layer to obtain the risk warning level, the method further includes: Obtaining the residual sum of squares of each target residual; Obtaining the risk contribution of each target medical indicator data according to each target residual and the residual square sum; The target risk contribution with the largest contribution is screened out from the risk contributions, and the target medical indicator data of the target risk contribution is used as the risk cause for the risk warning level.

8. A medical indicator risk warning device, characterized in that: The device comprises: A medical indicator acquisition module is used to collect target medical indicator data of target users; A model acquisition module is used to acquire a pre-trained risk warning model; wherein the risk warning model includes a coding layer, a decoding layer, a residual layer and a risk warning layer; A feature extraction module, used for extracting features of the target medical indicator data through the coding layer to obtain medical indicator features; A risk level identification module, used to identify the risk level of the medical indicator feature through the decoding layer to obtain the medical indicator risk level of the target medical indicator data; A residual calculation module, used for obtaining the observed medical indicator level of the target medical indicator data, and calculating the target residual between the medical indicator risk level and the observed medical indicator level through the residual layer; The risk warning module is used to perform risk warning on the target medical indicator data according to the target residual and the risk warning layer to obtain the risk warning level.

9. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the medical indicator risk warning method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the medical indicator risk warning method according to any one of claims 1 to 7 is implemented.

Images