Behavior early warning method and device for chronic disease patient, electronic equipment and storage medium

Abstract

The embodiment of the invention relates to a behavior early warning method and device for a chronic disease patient, electronic equipment and a storage medium, and the method comprises the steps: obtaining the physiological index variance and behavior index variance of the chronic disease patient in a plurality of continuous time periods, and enabling the plurality of continuous time periods to comprise a current time period and a historical time period before the current time period; constructing the physiological index variances and the behavior index variances of a plurality of continuous time periods into multi-dimensional feature vectors arranged according to a time sequence; inputting the multi-dimensional feature vector into a trained LSTM model, and outputting a physiological index predicted value of the chronic disease patient in a future time period; outputting early warning information under the condition that the predicted value of the physiological index exceeds a preset physiological threshold value and the variance of the physiological index and the variance of the behavior index in the current time period exceed corresponding preset variance threshold values; according to the method, the physiological indexes of the patient are predicted through the LSTM model, and when the physiological indexes, the physiological index variance and the behavior index variance of the patient exceed threshold values, behavior early warning information can be output.

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and particularly to a method, device, electronic device and storage medium for behavior warning of chronic disease patients. Background Art

[0002] Chronic diseases (such as hypertension, diabetes, hyperlipidemia, etc.) are extremely common in modern society and require long-term monitoring and treatment. The daily behaviors of patients, such as reasonable diet, regular exercise, etc., play a key role in disease control.

[0003] However, most traditional methods rely on intermittent medical examination data, which are difficult to effectively capture the behavior patterns of patients and cannot achieve real-time and accurate prediction of the behavior changes and health trends of patients. Summary of the Invention

[0004] The present invention provides a method, device, electronic device and storage medium for behavior warning of chronic disease patients to solve the technical problem that real-time and accurate prediction of the behavior changes and health trends of patients cannot be achieved.

[0005] In a first aspect, the present invention provides a method for behavior warning of chronic disease patients, including: obtaining the variance of physiological indicators and the variance of behavior indicators of chronic disease patients in a plurality of consecutive time periods, where the plurality of consecutive time periods include the current time period and historical time periods before the current time period; constructing the variance of physiological indicators and the variance of behavior indicators in a plurality of consecutive time periods into a multi-dimensional feature vector arranged in chronological order; inputting the multi-dimensional feature vector into a trained LSTM model to output the predicted value of physiological indicators of chronic disease patients in a future time period; and outputting a warning message when the predicted value of physiological indicators exceeds a preset physiological threshold and the variance of physiological indicators and the variance of behavior indicators in the current time period exceed the corresponding preset variance thresholds.

[0006] In some embodiments, the obtaining the variance of physiological indicators and the variance of behavior indicators of chronic disease patients in a plurality of consecutive time periods, where the plurality of consecutive time periods include the current time period and historical time periods before the current time period, includes: obtaining the physiological indicator values and behavior indicator values corresponding to fixed time points in the current time period and historical time periods of chronic disease patients; and calculating the variance of physiological indicators corresponding to each physiological indicator value and the variance of behavior indicators corresponding to each behavior indicator value by using a sliding time window method.

[0007] In some embodiments, the method further includes: obtaining patient individual information and performing one-hot encoding on the patient individual information; and constructing the one-hot encoded patient individual information, the variance of physiological indicators and the variance of behavior indicators in a plurality of consecutive time periods into a multi-dimensional feature vector arranged in chronological order.

[0008] In some embodiments, the physiological indicators include at least one of the following: blood glucose, blood pressure, blood lipid; the behavioral indicators include at least one of the following: diet, exercise, medication; the patient individual information includes at least one of the following: age, gender, weight, medical history.

[0009] In some embodiments, the method further includes: finding out an intervention measure that matches the predicted value of the physiological indicators of the chronic disease patient, the variance of the physiological indicators corresponding to the current time period, and the variance of the behavioral indicators from a preset rule base.

[0010] In a second aspect, the present invention provides a behavioral warning device for a chronic disease patient, including: an acquisition module, configured to acquire the variance of physiological indicators and the variance of behavioral indicators of a chronic disease patient in a plurality of consecutive time periods, where the plurality of consecutive time periods include the current time period and historical time periods before the current time period; a construction module, configured to construct the variance of physiological indicators and the variance of behavioral indicators in a plurality of consecutive time periods into a multi-dimensional feature vector arranged in chronological order; a prediction module, configured to input the multi-dimensional feature vector into a trained LSTM model and output a predicted value of the physiological indicators of the chronic disease patient in a future time period; a warning module, configured to output a warning message when the predicted value of the physiological indicators exceeds a preset physiological threshold and the variance of the physiological indicators and the variance of the behavioral indicators in the current time period exceed corresponding preset variance thresholds.

[0011] In some embodiments, the acquisition module is specifically configured to: acquire the values of physiological indicators and behavioral indicators corresponding to fixed time points in the current time period and historical time periods of a chronic disease patient; calculate the variance of the physiological indicators corresponding to each physiological indicator value and the variance of the behavioral indicators corresponding to each behavioral indicator value by using a sliding time window method.

[0012] In some embodiments, the acquisition module is further configured to: acquire patient individual information and perform one-hot encoding on the patient individual information; the construction module is specifically configured to: construct the one-hot encoded patient individual information, the variance of the physiological indicators, and the variance of the behavioral indicators in a plurality of consecutive time periods into a multi-dimensional feature vector arranged in chronological order.

[0013] In a third aspect, the present invention provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, where the processor, the communication interface, and the memory complete communication with each other through the communication bus; the memory is used to store a computer program; when the processor executes the program stored on the memory, it implements the steps of the behavioral warning method for a chronic disease patient according to any one of the first aspects.

[0014] Fourthly, the present invention provides a computer-readable storage medium, on which a computer program is stored, characterized in that when the computer program is executed by a processor, the steps of the behavior warning method for chronic disease patients according to any one of the first aspect are realized.

[0015] A behavior warning method, device, electronic device and storage medium for chronic disease patients provided by an embodiment of the present invention can learn the relationship between the physiological indicators of chronic disease patients and the patients' behaviors through an LSTM model, so as to accurately predict the physiological indicators of the patients. When the predicted physiological indicators of the patients exceed the normal values and the variance of the physiological indicators exceeds a certain threshold, if the recent behaviors of the patients fluctuate greatly, it means that the recent behaviors of the patients have a great impact on the patients' health, and warning information for adjusting the patients' behaviors (diet, exercise amount) can be output. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings here are incorporated into the specification and form a part of this specification, showing the embodiments consistent with the present invention and used together with the specification to explain the principles of the present invention.

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, other drawings can be obtained according to these drawings without creative efforts.

[0018] Figure 1 It is a schematic flowchart of a behavior warning method for chronic disease patients provided by an embodiment of the present invention;

[0019] Figure 2 It is a schematic flowchart of another behavior warning method for chronic disease patients provided by an embodiment of the present invention;

[0020] Figure 3 It is a schematic flowchart of still another behavior warning method for chronic disease patients provided by an embodiment of the present invention;

[0021] Figure 4 It is a schematic structural diagram of a behavior warning device for chronic disease patients provided by an embodiment of the present invention;

[0022] Figure 5 It is a schematic hardware structure diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0024] Figure 1 This is a flowchart showing a method for behavior warning of chronic disease patients provided by an embodiment of the present invention. As Figure 1 shown, the method includes the following steps:

[0025] Step S101: Obtain the variance of physiological indicators and the variance of behavior indicators of a chronic disease patient in a plurality of consecutive time periods, where the plurality of consecutive time periods include the current time period and historical time periods before the current time period.

[0026] Specifically, the physiological indicators include at least one of the following: blood glucose, blood pressure, and blood lipid; the behavior indicators include at least one of the following: diet, exercise, and medication; the time period can be understood as a fixed time step, such as 1 day. In this step, obtain the variance of blood glucose, blood pressure, and blood lipid, as well as the variance of diet, exercise, and medication of a chronic disease patient for multiple consecutive days. For example, the variance of blood glucose reflects the blood glucose fluctuation of the patient within a certain period of time, and the variance of exercise volume and diet reflects the stability of the patient's living habits within a certain period of time.

[0027] In some embodiments, step S101 includes: obtaining the physiological indicator values and behavior indicator values corresponding to fixed time points in the current time period and historical time periods of a chronic disease patient; calculating the variance of physiological indicators corresponding to each physiological indicator value and the variance of behavior indicators corresponding to each behavior indicator value by using a sliding time window method.

[0028] Specifically, continuously collect physiological indicator data such as blood pressure, blood glucose, and blood lipid of a chronic disease patient at fixed time points every day (for example). For blood pressure data, focus on the fluctuation of blood pressure before taking medicine every day. For blood glucose data, according to the meal time and drug action time, focus on fasting and two hours after meals. For blood lipid data, since its fluctuation is relatively slow, based on the detection data once a month, the data for 3 consecutive months form an analysis interval; at the same time, continuously collect behavior indicator data such as the diet, exercise, and medication of the patient. Among them, the dietary intake can be quantified as the weight of food, etc. For example, record the weight of carbohydrates ingested daily, and the exercise volume can also be quantified in various ways, such as using devices such as sports bracelets to record data such as the number of steps and exercise duration.

[0029] After the data is collected, preliminary verification and preprocessing will be performed on the data (such as removing outliers, filling in missing values, etc.), and the data will be standardized, such as normalizing it to the interval [0, 1], to achieve the unification of data formats and standards, facilitating subsequent data integration.

[0030] Then, use the variance formula (1) to calculate the variance of each indicator:

[0031]

[0032] where s2 represents the variance, n is the number of corresponding indicators, and x i is the i-th data of the corresponding indicator, and x is the mean of the corresponding indicator.

[0033] For example, for the blood glucose variance, obtain the fasting blood glucose values a1, a2,..., a7 for 7 consecutive days, and calculate the average value Then calculate the fasting blood glucose variance according to the variance formula (1); obtain the blood glucose values 2 hours after meals b1, b2,..., b7 for 7 consecutive days, and calculate the average value Then calculate the blood glucose variance 2 hours after meals according to the variance formula (1).

[0034] For the exercise variance, taking the number of steps as the quantification index of exercise volume, obtain the number of exercise steps y1, y2,..., y7 for 7 consecutive days, and first calculate the average number of steps Then use the variance formula (1) to calculate the exercise variance.

[0035] For the diet volume variance, taking the weight of carbohydrates ingested daily as the quantification index, obtain the weights of carbohydrates ingested z1, z2,..., z7 for 7 consecutive days, and calculate the average weight Then use the variance formula (1) to calculate the diet variance.

[0036] Assume that the input of the LSTM model is the variance data of each dimension for 7 consecutive days. Then the physiological index values for the 1st - 10th days can be obtained. The time dimension of the sliding window is 4 days, so the sliding window can be moved to the 1st - 4th days. Substitute the physiological index values of these 4 days into formula (1) to calculate the physiological index variance corresponding to the 4th day. Then move the sliding window to the 2nd - 5th days and calculate the physiological index variance corresponding to the 5th day, and so on, calculating the physiological index variances for the 6th, 7th, 8th, 9th, and 10th days, thereby obtaining the physiological index variances for 7 consecutive days (i.e., the 4th - 10th days). Similarly, the behavioral index variances for 7 consecutive days can be obtained.

[0037] Step S102: Construct a multi-dimensional feature vector arranged in chronological order from the physiological index variances and behavioral index variances for multiple consecutive time periods.

[0038] Specifically, variance data of multiple dimensions over consecutive days are constructed into a multi-dimensional feature vector arranged in chronological order. For example, the variance of physiological indicators and the variance of behavioral indicators over 7 consecutive days are arranged in chronological order to form a multi-dimensional feature vector.

[0039] Step S103: Input the multi-dimensional feature vector into the trained LSTM model, and output the predicted value of the physiological indicators of the chronic disease patient in the future time period.

[0040] Specifically, the long short-term memory (LSTM) model includes an input layer, a hidden layer, and an output layer; the integrated multi-dimensional feature vector is used as the input of the LSTM model. The hidden layer learns and extracts the long-term dependence relationship in the time series data through LSTM units, captures the pattern of data changing over time, and the output layer is used to predict the physiological indicator value of the patient for the next day.

[0041] It should be noted that before applying the LSTM model, the LSTM model needs to be trained first using a large amount of existing chronic disease patient data. During the training process, the mean squared error loss function is used to calculate the average of the sum of the squares of the differences between the predicted value and the true value to measure the deviation degree between the model's predicted value and the true value. By minimizing the mean squared error loss function, the model can better fit the data, reduce the error between the predicted value and the true value, and improve the prediction accuracy of the model. The formula for the mean squared error loss function is as follows:

[0042]

[0043] where m is the number of training samples, y j is the true physiological indicator value, is the predicted value of the physiological indicator.

[0044] Step S104: When the predicted value of the physiological indicator exceeds the preset physiological threshold, and the variance of the physiological indicator and the variance of the behavioral indicator in the current time period exceed the corresponding preset variance thresholds, an early warning message is output.

[0045] Specifically, after inputting the current multi-dimensional feature vector of a chronic disease patient into the trained LSTM model, the predicted values of physiological indicators for a future time period are output. If the predicted value of this physiological indicator exceeds a preset physiological threshold, an analysis will be carried out in combination with the variance of physiological indicators, the variance of exercise, and the variance of diet on the same day or in the recent period. If the variance of physiological indicators and the variance of behavioral indicators also exceed the preset variance threshold, then it is determined that the patient has abnormal behavior. For example, when the blood sugar predicted by the LSTM model exceeds the preset value (such as fasting blood sugar higher than 6.1 mmol / L and blood sugar 2 hours after a meal higher than 7.8 mmol / L), and the variance of blood sugar exceeds the preset value, and the variance of diet also exceeds the preset value, it indicates that the abnormal blood sugar is caused by an unreasonable diet structure of the patient; if the variance of the patient's blood sugar exceeds the preset value and the variance of exercise exceeds the preset value, it means that the change in the patient's exercise habit affects the blood sugar fluctuation; at this time, a warning message should be sent in a timely manner.

[0046] In some embodiments, the method further includes: searching for an intervention measure matching the predicted value of the physiological indicator of the chronic disease patient, the variance of the physiological indicator corresponding to the current time period, and the variance of the behavioral indicator from a preset rule library.

[0047] Specifically, a rule library based on medical expert experience and clinical research results is established in advance. The rule library contains adjustment plans for different abnormal behaviors and disease conditions. After the LSTM model predicts that the patient has abnormal behavior, according to the type of abnormality and the specific condition information of the patient, the corresponding adjustment plan is matched from the rule library. For example, if it is determined that the patient's blood sugar is abnormally elevated due to irregular diet, the corresponding adjustment plan in the rule library may include providing detailed diet advice. In addition, the doctor can further adjust and optimize the adjustment plan according to the specific situation of the patient, such as whether there are other complications and the current physical condition.

[0048] The behavior warning method for chronic disease patients provided in this embodiment can learn the relationship between the physiological indicators of chronic disease patients and the patients' behaviors through the LSTM model, so as to accurately predict the physiological indicators of the patients. When the predicted physiological indicators of the patients exceed the normal value and the variance of the physiological indicators exceeds a certain threshold, if the recent behavior fluctuations of the patients are large, it means that the recent behaviors of the patients have a great impact on the patients' health, and a warning message for adjusting the patients' behaviors (diet, exercise amount) can be output.

[0049] On the basis of the foregoing embodiments, Figure 2 is a schematic flowchart of another behavior warning method for chronic disease patients provided by the embodiment of the present invention. As Figure 2 shown, it includes the following steps:

[0050] Step S201: Obtain the variance of physiological indicators and the variance of behavioral indicators of a chronic disease patient in multiple consecutive time periods, where the multiple consecutive time periods include the current time period and historical time periods before the current time period.

[0051] Step S202: Obtain the patient's individual information and perform one-hot encoding on the patient's individual information.

[0052] Step S203: Construct a multi-dimensional feature vector arranged in chronological order from the one-hot encoded patient's individual information, the variance of physiological indicators, and the variance of behavioral indicators in multiple consecutive time periods.

[0053] Step S204: Input the multi-dimensional feature vector into a trained LSTM model to output the predicted value of the physiological indicators of the chronic disease patient in the future time period.

[0054] Step S205: Output a warning message when the predicted value of the physiological indicators exceeds a preset physiological threshold and the variances of the physiological indicators and behavioral indicators in the current time period exceed the corresponding preset variance thresholds.

[0055] It should be noted that the implementation manners of steps S201, S204, and S205 in this embodiment are similar to those of steps S101, S103, and S104 in the foregoing embodiment, and will not be elaborated here.

[0056] The difference from the foregoing embodiment is that, in order to further improve the accuracy of health prediction and behavior warning for different individuals, in this embodiment, the patient's individual information is obtained and one-hot encoded; the one-hot encoded patient's individual information, the variance of physiological indicators, and the variance of behavioral indicators in multiple consecutive time periods are constructed into a multi-dimensional feature vector arranged in chronological order.

[0057] Specifically, in the data collection stage, in addition to conventional blood glucose, exercise, and diet data, information reflecting individual differences such as age, gender, weight, height, genetic history, and complication status needs to be collected. Taking diabetic patients as an example, the blood glucose metabolism mechanisms of adolescent patients and elderly patients are different. The former has relatively strong metabolism, while the latter may have problems with metabolic function decline. Integrating this information into the input data of the LSTM model provides a basis for the model to learn individual difference features.

[0058] Normalize and standardize the data of different individuals to unify the data scale. For data such as blood glucose values, exercise amounts, and dietary intakes, the value ranges of different individuals vary greatly. Use the Z-score standardization or Min-Max normalization method to map the data to a specific interval; perform one-hot encoding on the categorical information (such as gender, medical history, etc.) in the patient's basic information, so as to ensure that when the LSTM model is trained, the data of different individuals can be compared and learned on the same scale, and avoid model learning biases caused by data scale differences.

[0059] Based on the foregoing embodiments, by obtaining the patient's individual information and performing one-hot encoding on the patient's individual information; constructing a multi-dimensional feature vector arranged in chronological order from the one-hot encoded patient's individual information and the variances of physiological indicators and behavioral indicators for multiple consecutive time periods, so that the LSTM model can more accurately adapt to different individuals and improve the prediction effect.

[0060] To further understand the present invention, Figure 3 is a schematic flowchart of another behavior warning method for chronic disease patients provided by an embodiment of the present invention, as Figure 3 shown, including the following steps:

[0061] (1) Collection of physiological indicators, behavioral indicators, and patient individual information: Continuously collect the patient's blood pressure, blood glucose, blood lipids, exercise amount (number of steps, duration), dietary intake (weight of carbohydrates, etc.), medication records, and the patient's individual information, including age, gender, weight, medical history, etc.

[0062] (2) Data preprocessing: Remove outliers, fill in missing values, perform standardization processing (such as normalizing to the [0,1] interval), and perform one-hot encoding on categorical data (gender, medical history).

[0063] (3) Variance calculation: Calculate the variances of data such as blood glucose, blood pressure, blood lipids, exercise amount, and dietary intake within a fixed interval respectively.

[0064] (4) Construct a multi-dimensional feature vector: Integrate the variance data of each dimension (blood glucose, blood pressure, exercise, diet, etc.) with the patient's basic information (age, gender, medical history) to form a time series input, and the input structure example is as follows:

[0065] Date, blood glucose variance, blood pressure variance, exercise variance, diet variance, age, gender encoding, medical history encoding.

[0066] (5) LSTM model construction and training: Determine the initial LSTM model. A large amount of existing chronic disease patient data can be divided into a training set and a test set, and the LSTM model is trained using the training set to minimize the value of the mean square error loss function.

[0067] (6) LSTM model evaluation: Use the test set to evaluate the trained LSTM model. If the prediction accuracy of the LSTM model is relatively high, then execute step (7); otherwise, return to execute (5).

[0068] (7) Save the trained LSTM model.

[0069] (8) Input the multi-dimensional feature vector of the new patient into the trained LSTM model to predict the physiological index prediction value in the future time period.

[0070] (9) Abnormality judgment: If the physiological index prediction value exceeds the preset threshold, and the variances of the physiological index and the behavior index exceed the preset threshold, it is determined that the patient's behavior is abnormal, such as abnormal diet, abnormal exercise volume, etc.; if not, return to (8) to continue monitoring.

[0071] (10) Match the corresponding intervention measures from the preset rule library and send them to the doctor.

[0072] (11) After the doctor optimizes the intervention measures, remind the patient to change the diet or exercise volume.

[0073] In summary, in this embodiment, through the combination of variance calculation and the LSTM model, the dual optimization of "data volatility analysis + time series prediction" is realized, that is, variance calculation is performed on data such as blood glucose, blood pressure, blood lipid, exercise volume, and dietary intake of chronic disease patients at fixed time points in continuous time, and the LSTM model is used for in-depth analysis to realize real-time and accurate prediction of the patient's behavior changes and health trends. When there is an abnormality, timely reminder and early warning are given to the doctor, allowing the doctor to intervene and follow up, and ultimately affecting the patient's behavior.

[0074] Figure 4 It is a schematic structural diagram of a behavior warning device for chronic disease patients provided by an embodiment of the present invention. As Figure 4 shown, the device includes:

[0075] An acquisition module 401, configured to acquire the variances of physiological indexes and behavior indexes of chronic disease patients in a plurality of consecutive time periods, where the plurality of consecutive time periods include the current time period and historical time periods before the current time period;

[0076] A construction module 402, configured to construct the variances of physiological indexes and behavior indexes in a plurality of consecutive time periods into a multi-dimensional feature vector arranged in chronological order;

[0077] A prediction module 403, configured to input the multi-dimensional feature vector into the trained LSTM model and output the physiological index prediction value of chronic disease patients in the future time period;

[0078] The warning module 404 is used to output a warning message when the predicted value of the physiological index exceeds the preset physiological threshold, and the variances of the physiological index and the behavior index in the current time period exceed the corresponding preset variance thresholds.

[0079] In some embodiments, the obtaining module 401 is specifically configured to:

[0080] Obtain the physiological index values and behavior index values corresponding to the fixed time points of the chronic disease patient in the current time period and the historical time period;

[0081] Calculate the variance of the physiological index corresponding to each physiological index value and the variance of the behavior index corresponding to each behavior index value by using a sliding time window method.

[0082] In some embodiments, the obtaining module 401 is further configured to:

[0083] Obtain the patient individual information and perform one-hot encoding on the patient individual information;

[0084] The constructing module 402 is specifically configured to: construct the one-hot encoded patient individual information, as well as the variances of the physiological index and the behavior index in multiple consecutive time periods, into a multi-dimensional feature vector arranged in chronological order.

[0085] In some embodiments, the physiological index includes at least one of the following: blood glucose, blood pressure, and blood lipid;

[0086] The behavior index includes at least one of the following: diet, exercise, and medication;

[0087] The patient individual information includes at least one of the following: age, gender, weight, and medical history.

[0088] In some embodiments, the warning module 404 is further configured to:

[0089] Search for an intervention measure that matches the predicted value of the physiological index of the chronic disease patient, the variance of the physiological index in the current time period, and the variance of the behavior index from a preset rule library.

[0090] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working process and corresponding beneficial effects of the above-described behavior warning device for chronic disease patients can refer to the corresponding process in the foregoing method examples, and will not be elaborated here.

[0091] Figure 5 This is a schematic hardware structure diagram of an electronic device provided by an embodiment of the present invention, as Figure 5As shown, the electronic device includes: a processor 501, a communication interface 502, a memory 503, and a communication bus 504. Among them, the processor 501, the communication interface 502, and the memory 503 communicate with each other through the communication bus 504.

[0092] The memory 503 is used to store computer programs.

[0093] In an embodiment of the present application, when the processor 501 is used to execute the program stored on the memory 503, it implements the steps of the behavior warning method for chronic disease patients provided by any of the foregoing method embodiments.

[0094] The electronic device provided by the embodiment of the present application has the same implementation principle and technical effects as the above embodiment, and will not be elaborated here.

[0095] The above memory 503 can be an electronic memory such as flash memory, EEPROM (electrically erasable programmable read-only memory), EPROM, hard disk, or ROM. The memory 503 has a storage space for program codes for executing any method steps in the above methods. For example, the storage space for program codes can include respective program codes for implementing each step in the above methods. These program codes can be read from or written into one or more computer program products. These computer program products include program code carriers such as hard disks, optical discs (CDs), memory cards, or floppy disks. Such computer program products are usually portable or fixed storage units. The storage unit can have a storage segment or storage space arranged similarly to the memory 503 in the above electronic device. The program codes can be compressed in an appropriate form, for example. Generally, the storage unit includes a program for executing the method steps according to the embodiments of the present application, that is, codes that can be read by a processor such as 501, and when these codes are run by the electronic device, they cause the electronic device to execute each step in the method described above.

[0096] An embodiment of the present application also provides a computer-readable storage medium. A computer program is stored on the above computer-readable storage medium, and when the computer program is executed by a processor, it implements the steps of the behavior warning method for chronic disease patients as described above.

[0097] The above computer-readable storage medium can be included in the device / device described in the above embodiment; it can also exist separately without being assembled into the device / device. The above computer-readable storage medium carries one or more programs, and when the one or more programs are executed, they implement the method according to the embodiments of the present application.

[0098] According to an embodiment of the present application, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, it may include but is not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In the present application, the computer-readable storage medium may be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, apparatus, or device.

[0099] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the said element.

[0100] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather to the broadest scope consistent with the principles and novel features claimed herein.

Claims

1. A behavior warning method for patients with chronic diseases, characterized in that, Including: Obtain the variance of physiological indicators and the variance of behavioral indicators of chronic disease patients in multiple consecutive time periods, where the multiple consecutive time periods include the current time period and historical time periods before the current time period; Construct a multi-dimensional feature vector arranged in chronological order from the variance of physiological indicators and the variance of behavioral indicators in multiple consecutive time periods; Input the multi-dimensional feature vector into a trained LSTM model to output the predicted value of the physiological indicators of chronic disease patients in the future time period; When the predicted value of the physiological indicators exceeds the preset physiological threshold, and the variance of the physiological indicators and the variance of the behavioral indicators in the current time period exceed the corresponding preset variance thresholds, output a warning message.

2. The method according to claim 1, wherein The obtaining the variance of physiological indicators and the variance of behavioral indicators of chronic disease patients in multiple consecutive time periods, where the multiple consecutive time periods include the current time period and historical time periods before the current time period, includes: Obtain the physiological indicator values and behavioral indicator values corresponding to fixed time points in the current time period and historical time periods of chronic disease patients; Calculate the variance of physiological indicators corresponding to each physiological indicator value and the variance of behavioral indicators corresponding to each behavioral indicator value by using a sliding time window method.

3. The method according to claim 1 or 2, characterized in that, The method further includes: Obtain patient individual information and perform one-hot encoding on the patient individual information; Construct a multi-dimensional feature vector arranged in chronological order from the one-hot encoded patient individual information, the variance of physiological indicators, and the variance of behavioral indicators in multiple consecutive time periods.

4. The method according to claim 3, wherein Physiological indicators include at least one of the following: blood glucose, blood pressure, blood lipid; Behavioral indicators include at least one of the following: diet, exercise, medication; Patient individual information includes at least one of the following: age, gender, weight, medical history.

5. The method according to claim 3, characterized in that, The method further includes: Search for intervention measures that match the predicted value of the physiological indicators of the chronic disease patient, the variance of the physiological indicators in the current time period, and the variance of the behavioral indicators from a preset rule library.

6. A behavior warning device for chronic disease patients, characterized in that, Including: An obtaining module for obtaining the variance of physiological indicators and the variance of behavioral indicators of chronic disease patients in multiple consecutive time periods, where the multiple consecutive time periods include the current time period and historical time periods before the current time period; A constructing module for constructing a multi-dimensional feature vector arranged in chronological order from the variance of physiological indicators and the variance of behavioral indicators in multiple consecutive time periods; A predicting module for inputting the multi-dimensional feature vector into a trained LSTM model to output the predicted value of the physiological indicators of chronic disease patients in the future time period; A warning module for outputting a warning message when the predicted value of the physiological indicators exceeds the preset physiological threshold, and the variance of the physiological indicators and the variance of the behavioral indicators in the current time period exceed the corresponding preset variance thresholds.

7. The device according to claim 6, characterized in that, The obtaining module is specifically used for: Obtain the physiological indicator values and behavioral indicator values corresponding to fixed time points in the current time period and historical time periods of chronic disease patients; Calculate the variance of physiological indicators corresponding to each physiological indicator value and the variance of behavioral indicators corresponding to each behavioral indicator value by using a sliding time window method.

8. The device according to claim 6 or 7, characterized in that, The obtaining module is further used for: Obtain patient individual information and perform one-hot encoding on the patient individual information; The building block is specifically used for: constructing a multi-dimensional feature vector arranged in chronological order from the one-hot encoded patient individual information and the variances of physiological indexes and behavior indexes in a plurality of consecutive time periods.

9. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete mutual communication through the communication bus; The memory is used for storing computer programs; When the processor is used to execute the program stored on the memory, it realizes the steps of the behavior warning method for chronic disease patients described in any one of claims 1-5.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it realizes the steps of the behavior warning method for chronic disease patients described in any one of claims 1-5.

Images