Clinical data analysis method and device, computer equipment and storage medium
By preprocessing clinical data and deep learning model analysis, the problem of inaccurate clinical data analysis in the existing technology is solved, efficient and accurate detection of NSTEMI patients is achieved, and the risk of missed diagnosis is reduced.
Patent Information
- Application Number
- CN202311344322.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-17
- Publication Date
- 2025-04-18
AI Technical Summary
The prior art is not accurate enough when analyzing clinical data, especially when detecting NSTEMI patients, which can easily lead to missed diagnosis.
By obtaining biochemical indexes and electrocardiogram data, after preprocessing, the data is analyzed using the deep learning model obtained based on the activation model, pooling model and fully connected model training, and the data is generated and vector analysis is performed to improve the analysis accuracy.
It realizes efficient and accurate analysis of clinical data, improves analysis accuracy, and reduces the risk of missed diagnosis in NSTEMI patients.
Smart Images
Figure CN119851838A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of clinical data analysis, and in particular to a clinical data analysis method, device, computer equipment and storage medium. Background Art
[0002] With the rapid development of computer microelectronics technology and signal processing technology, ECG signal acquisition equipment is also developing in the direction of intelligence and portability. At the same time, in actual clinical practice, according to the diagnosis and treatment guidelines proposed by the American College of Cardiology, the European Society of Cardiology, and the Chinese Medical Association, the diagnostic criteria for acute myocardial infarction (AMI) must have at least two of the following three criteria: first, clinical history of ischemic chest pain, second, dynamic evolution of electrocardiogram, and third, dynamic changes in the concentration of blood myocardial markers of myocardial necrosis. The standard points out that it is necessary to weaken the actual effect of the simple electrocardiogram (ECG)-assisted clinical physical examination diagnosis model.
[0003] Electrocardiogram is a conventional auxiliary diagnostic and treatment method for acute myocardial infarction. It has the advantages of low price, high safety and synchronous reporting, but it cannot be used alone as a complete sign to determine whether one has ischemic heart disease. The typical electrocardiogram of patients with myocardial infarction shows arched ST segment elevation or T wave inversion (TWI) and pathological Q wave (QW) in patients with old myocardial infarction. The disease that can find the above electrocardiogram changes is called ST segment elevation myocardial infarction (STEMI). However, STEMI with ST segment elevation only accounts for about 20% of myocardial infarction events, while more than 70% of myocardial infarction patients do not have the above conventional electrocardiogram manifestations, usually showing ST segment depression, T wave changes or no obvious changes in QRS complex. Such patients without obvious symptoms are called NSTEMI (Non ST-segment elevation myocardial infarction). However, there is currently no simple method to detect NSTEMI, and most NSTEMI patients have negative results in conventional electrocardiograms, which can easily lead to missed diagnosis.
[0004] With the rise and development of artificial intelligence technology, the analysis and processing of ECG data through artificial intelligence technology can support functions such as disease auxiliary diagnosis, health management, and remote consultation. However, the current effect of artificial intelligence technology on ECG data analysis is not good, which leads to poor diagnosis results based on the analysis results. Summary of the invention
[0005] Based on this, it is necessary to propose a clinical data analysis method, device, computer equipment and storage medium to address the technical problem that the existing technology is not accurate enough in analyzing clinical data.
[0006] In a first aspect, a clinical data analysis method is provided, the method comprising:
[0007] Acquiring clinical data, wherein the clinical data includes biochemical indicators and / or electrocardiogram data of the target user;
[0008] Preprocessing the clinical data to obtain target data;
[0009] Performing data analysis on the target data through a trained deep learning model to obtain an analysis vector, wherein the deep learning model is a model trained based on an activation model, a pooling model, and a fully connected model connected in sequence;
[0010] Vector analysis is performed according to the analysis vector to obtain an analysis result.
[0011] In a second aspect, a clinical data analysis device is provided, the device comprising:
[0012] An acquisition module, used for acquiring clinical data, wherein the clinical data includes biochemical indicators and / or electrocardiogram data of the target user;
[0013] A preprocessing module, used for preprocessing the clinical data to obtain target data;
[0014] A vector module, used to perform data analysis on the target data through a trained deep learning model to obtain an analysis vector, wherein the deep learning model is a model trained based on an activation model, a pooling model, and a fully connected model connected in sequence;
[0015] The analysis module is used to perform vector analysis according to the analysis vector to obtain an analysis result.
[0016] In a third aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned clinical data analysis method when executing the computer program.
[0017] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned clinical data analysis method are implemented.
[0018] The clinical data analysis method proposed in the present invention obtains clinical data, wherein the clinical data includes biochemical indicators and / or electrocardiogram data of the target user, and then preprocesses the clinical data to obtain target data, and performs data analysis on the target data through a trained deep learning model to obtain an analysis vector, wherein the deep learning model is a model trained based on an activation model, a pooling model, and a fully connected model connected in sequence, and finally performs vector analysis based on the analysis vector to obtain an analysis result. The deep learning model trained based on the activation model, the pooling model, and the fully connected model can realize efficient and accurate data analysis of the clinical data, thereby obtaining an analysis result with improved analysis accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0020] in:
[0021] Figure 1 is an application environment diagram of a clinical data analysis method in an embodiment;
[0022] Figure 2 is a flow chart of a clinical data analysis method in one embodiment;
[0023] Figure 3 is a structural block diagram of a clinical data analysis device in one embodiment;
[0024] Figure 4 is a structural block diagram of a computer device in one embodiment;
[0025] Figure 5 It is a structural block diagram of a computer device in another embodiment. DETAILED DESCRIPTION
[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0027] The clinical data analysis method provided by the embodiment of the present invention can be applied in Figure 1In an application environment, the client 110 communicates with the server 120 through a network. The server 120 can receive clinical data through the client 110, wherein the clinical data includes biochemical indicators and / or electrocardiogram data of the target user, and the server 120 pre-processes the clinical data to obtain the target data, and performs data analysis on the target data through a trained deep learning model to obtain an analysis vector, wherein the deep learning model is a model obtained by training an activation model, a pooling model, and a fully connected model connected in sequence, and finally the server 120 performs vector analysis according to the analysis vector to obtain an analysis result, and can realize efficient and accurate data analysis of clinical data through a deep learning model obtained by training an activation model, a pooling model, and a fully connected model, thereby obtaining an analysis result with improved analysis accuracy. Among them, the client 110 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, and portable wearable devices. The server 120 can be implemented with an independent server or a server cluster consisting of multiple servers. The present invention is described in detail below through specific embodiments.
[0028] See also Figure 2 As shown, Figure 2 A flow chart of a clinical data analysis method provided by an embodiment of the present invention includes the following steps:
[0029] Step S101: Acquire clinical data, wherein the clinical data includes biochemical indicators and / or electrocardiogram data of the target user;
[0030] The biochemical index may be serum myocardial enzymes, troponin T, brain natriuretic peptide (BNP), etc., and the electrocardiogram data may be electrocardiogram signals. As an example, the clinical data may be a physical examination report provided by the target user.
[0031] Step S102: preprocessing the clinical data to obtain target data;
[0032] For example, standardizing the names of biochemical indicators to obtain standardized biochemical indicators is conducive to improving the accuracy of subsequent data analysis. De-noising electrocardiogram data is performed to obtain de-noised data. The standardized biochemical indicators and de-noised data are used as target data.
[0033] Furthermore, in one embodiment, the step of preprocessing the clinical data to obtain target data includes:
[0034] Step S1021: using discrete wavelet packet transform and sliding average method to perform noise reduction processing on the electrocardiogram data to obtain noise-reduced data;
[0035] Step S1022: Use the Pan-Tompkins algorithm and wavelet packet transform to locate the characteristic wave group of the denoised data to obtain the target data.
[0036] In this embodiment, discrete wavelet packet transform and sliding average method are used to perform noise reduction on power frequency interference, baseline drift and electromyographic interference in the electrocardiogram data, and the PT algorithm and wavelet packet transform are further used to locate the characteristic wave group of the noise-reduced data obtained after noise reduction, wherein the characteristic wave group can be QRS wave group, P wave and T wave.
[0037] Step S103: performing data analysis on the target data through a trained deep learning model to obtain an analysis vector, wherein the deep learning model is a model trained based on an activation model, a pooling model, and a fully connected model connected in sequence;
[0038] It should be noted that the deep learning model is a model trained based on the activation model, pooling model and fully connected model connected in sequence. The activation model is used to activate the target data, the pooling model is used to strengthen the features, and the fully connected model is used to connect the features.
[0039] Optionally, the deep learning model may be a medical data analysis model, wherein the medical data analysis model may be used to analyze medical data to obtain analysis text, and the medical data may be clinical data.
[0040] Optionally, the deep learning model may be a medical image generation model, wherein the medical image generation model is used to analyze clinical data and generate medical images corresponding to the clinical data.
[0041] Optionally, the deep learning model can be a medical data classification model, wherein the medical data classification model can be used to classify clinical data and generate one or more medical classification labels corresponding to the clinical data.
[0042] Furthermore, in one embodiment, the activation model includes a first convolutional layer, an activation layer, and a first pooling layer connected in sequence, and the step of performing data analysis on the target data using the trained deep learning model to obtain an analysis vector includes:
[0043] Step A, inputting the target data into a first convolutional layer for convolution processing, and obtaining a first processing result output by the first convolutional layer;
[0044] The first convolutional layer is a convolutional layer. In one implementation, the size of the first convolutional layer is 196*128, the step size is 6, and the number of filters in the first convolutional layer is 96.
[0045] Step B, inputting the first convolution result into an activation layer for activation processing, and inputting the activation processing result output by the activation layer into a first pooling layer for pooling processing to obtain a second processing result;
[0046] The activation layer uses the ReLU function for activation processing. The first pooling layer is a pooling layer. In one implementation, the size of the first pooling layer is 3×3, and the step size is 2. The first pooling layer can reduce the dimension of the activation processing result and retain the effective information.
[0047] Step C: obtaining an analysis vector according to the second processing result, the pooling model and the fully connected model.
[0048] In this embodiment, the second processing result is processed by the pooling model and the fully connected model to obtain an analysis vector.
[0049] Further, in one embodiment, the pooling model includes a first pooling module and a second pooling module connected in sequence, the first pooling module includes a second convolutional layer and a second pooling layer connected in sequence, and the step of obtaining the analysis vector according to the second processing result, the pooling model and the fully connected model includes:
[0050] Step C1, inputting the second processing result into the second convolutional layer for convolution processing, to obtain a third processing result output by the second convolutional layer;
[0051] The second convolutional layer is a convolutional layer. In one implementation, the size of the second convolutional layer is 5×5, and the number of filters in the second convolutional layer is 256.
[0052] Step C2, inputting the third processing result into the second pooling layer for pooling processing, to obtain a fourth processing result output by the second pooling layer;
[0053] The second pooling layer is a pooling layer. In one implementation, the size of the second pooling layer is 3×3 and the step size is 2.
[0054] Step C3: obtaining the analysis vector according to the fourth processing result, the second pooling module and the fully connected model.
[0055] Further, in one embodiment, the second pooling module includes a third convolutional layer, a fourth convolutional layer, a fifth convolutional layer and a third pooling layer connected in sequence, and the step of obtaining the analysis vector according to the fourth processing result, the second pooling module and the fully connected model includes:
[0056] Step C31, inputting the fourth processing result into the third convolutional layer for convolution processing, to obtain a fifth processing result output by the third convolutional layer;
[0057] Step C32, inputting the fifth processing result into the fourth convolutional layer for convolution processing, to obtain a sixth processing result output by the fourth convolutional layer;
[0058] Step C33, inputting the sixth processing result into the fifth convolutional layer for convolution processing, to obtain a seventh processing result output by the fifth convolutional layer;
[0059] Step C34, inputting the seventh processing result into the third pooling layer for pooling processing, to obtain an eighth processing result output by the third pooling layer;
[0060] Step C35: Obtain the analysis vector according to the eighth processing result and the fully connected model.
[0061] The third convolutional layer, the fourth convolutional layer, and the fifth convolutional layer all refer to a type of convolutional layer. In one implementation, the channel sizes of the third convolutional layer, the fourth convolutional layer, and the fifth convolutional layer are all 3*3, and the number of filters is 384, 384, and 256, respectively. The third pooling layer refers to a type of pooling layer.
[0062] Furthermore, in one embodiment, the fully connected model includes a first fully connected layer, a first Dropout layer, a second fully connected layer, and a second Dropout layer connected in sequence, and the step of obtaining the analysis vector according to the eighth processing result and the fully connected model includes:
[0063] Step C351, inputting the eighth processing result into the first fully connected layer for connection, to obtain a first connection result output by the first fully connected layer;
[0064] Step C352, inputting the first connection result into the first Dropout layer for anti-overfitting processing to obtain a target result;
[0065] Step C353, inputting the target result into the second fully connected layer for connection, and obtaining a second connection result output by the second fully connected layer;
[0066] Step C354: input the second connection result into the second Dropout layer for anti-overfitting processing to obtain the analysis vector.
[0067] Among them, the first fully connected layer and the second fully connected layer both refer to a type of fully connected layer. For example, the channels of the first fully connected layer and the second fully connected layer are both 4096.
[0068] The first Dropout layer and the second Dropout layer are both Dropout layers used to prevent overfitting of deep learning models.
[0069] Step S104: performing vector analysis according to the analysis vector to obtain an analysis result.
[0070] In one implementation, the symptom portrait corresponding to the clinical data can be generated by performing vector analysis on the analysis vector; or, the symptom analysis text corresponding to the clinical data can be generated by performing vector analysis on the analysis vector. In another implementation, the probability value can be generated by performing vector analysis on the analysis vector, wherein the probability value refers to the probability value of the target user suffering from myocardial infarction.
[0071] As an example, the biochemical indicators and ECG data of the target user are photographed through the camera module; the posture sensing module is used for motion processing to reduce the impact of the settings and the drift of the sensor, and reduce the data collection and transmission when the patient is in a supine position. Then the biochemical indicators and ECG data of the target user are transmitted to the Bluetooth module. The target user's ECG data is transmitted to the preset database using Bluetooth communication, and the ECG data in the database is amplified, denoised, and filtered. Then, the physiological signal analysis tool is used to convert the file format of the ECG data after amplification, denoising, and filtering, and batch import. The trained deep learning model is used on the cloud to analyze the batch imported ECG data.
[0072] The clinical data analysis method proposed in this embodiment obtains clinical data, wherein the clinical data includes biochemical indicators and / or electrocardiogram data of the target user, and then preprocesses the clinical data to obtain target data, and performs data analysis on the target data through a trained deep learning model to obtain an analysis vector, wherein the deep learning model is a model trained based on an activation model, a pooling model, and a fully connected model connected in sequence, and finally performs vector analysis based on the analysis vector to obtain an analysis result. The deep learning model trained based on the activation model, the pooling model, and the fully connected model can realize efficient and accurate data analysis of the clinical data, thereby obtaining an analysis result with improved analysis accuracy.
[0073] See also Figure 3 As shown, in one embodiment, a clinical data analysis device is provided, the device comprising: an acquisition module 10, for acquiring clinical data, wherein the clinical data comprises biochemical indicators and / or electrocardiogram data of a target user;
[0074] A preprocessing module 20, used for preprocessing the clinical data to obtain target data;
[0075] A vector module 30, configured to perform data analysis on the target data through a trained deep learning model to obtain an analysis vector, wherein the deep learning model is a model trained based on an activation model, a pooling model, and a fully connected model connected in sequence;
[0076] The analysis module 40 is used to perform vector analysis according to the analysis vector to obtain an analysis result.
[0077] The clinical data analysis method proposed in this embodiment obtains clinical data, wherein the clinical data includes biochemical indicators and / or electrocardiogram data of the target user, and then preprocesses the clinical data to obtain target data, and performs data analysis on the target data through a trained deep learning model to obtain an analysis vector, wherein the deep learning model is a model trained based on an activation model, a pooling model, and a fully connected model connected in sequence, and finally performs vector analysis based on the analysis vector to obtain an analysis result. The deep learning model trained based on the activation model, the pooling model, and the fully connected model can realize efficient and accurate data analysis of the clinical data, thereby obtaining an analysis result with improved analysis accuracy.
[0078] Furthermore, in one embodiment, the vector module 30 is further configured to:
[0079] Inputting the target data into a first convolutional layer for convolution processing, and obtaining a first processing result output by the first convolutional layer;
[0080] Inputting the first convolution result into the activation layer for activation processing, and inputting the activation processing result output by the activation layer into the first pooling layer for pooling processing to obtain a second processing result;
[0081] An analysis vector is obtained according to the second processing result, the pooling model and the fully connected model.
[0082] Furthermore, in one embodiment, the vector module 30 is further used to: input the second processing result into the second convolution layer for convolution processing to obtain a third processing result output by the second convolution layer;
[0083] Inputting the third processing result into the second pooling layer for pooling processing to obtain a fourth processing result output by the second pooling layer;
[0084] The analysis vector is obtained according to the fourth processing result, the second pooling module and the fully connected model.
[0085] Furthermore, in one embodiment, the vector module 30 is further used to: input the fourth processing result into the third convolution layer for convolution processing to obtain a fifth processing result output by the third convolution layer;
[0086] Inputting the fifth processing result into the fourth convolutional layer for convolution processing to obtain a sixth processing result output by the fourth convolutional layer;
[0087] Inputting the sixth processing result into the fifth convolutional layer for convolution processing to obtain a seventh processing result output by the fifth convolutional layer;
[0088] Inputting the seventh processing result into the third pooling layer for pooling processing to obtain an eighth processing result output by the third pooling layer;
[0089] The analysis vector is obtained according to the eighth processing result and the fully connected model.
[0090] Furthermore, in one embodiment, the vector module 30 is further used to: input the eighth processing result into the first fully connected layer for connection, to obtain a first connection result output by the first fully connected layer;
[0091] Inputting the first connection result into the first Dropout layer for anti-overfitting processing to obtain a target result;
[0092] Inputting the target result into the second fully connected layer for connection, and obtaining a second connection result output by the second fully connected layer;
[0093] The second connection result is input into the second Dropout layer for anti-overfitting processing to obtain the analysis vector.
[0094] Furthermore, in one embodiment, the preprocessing module 20 is further used to: use discrete wavelet packet transform and sliding average method to perform noise reduction processing on the electrocardiogram data to obtain noise-reduced data;
[0095] The Pan-Tompkins algorithm and wavelet packet transform are used to locate the characteristic wave groups of the denoised data and obtain the target data.
[0096] Furthermore, in one embodiment, the analysis module 40 is also used to: generate a symptom portrait corresponding to the clinical data by performing vector analysis on the analysis vector; or generate a symptom analysis text corresponding to the clinical data by performing vector analysis on the analysis vector.
[0097] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, it implements a function or step on the server side of a clinical data analysis method.
[0098] In one embodiment, a computer device is provided. The computer device may be a client, and its internal structure diagram may be as follows: Figure 5 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it implements the functions or steps on the client side of a clinical data analysis method.
[0099] In one embodiment, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following steps are implemented:
[0100] Acquiring clinical data, wherein the clinical data includes biochemical indicators and / or electrocardiogram data of the target user;
[0101] Preprocessing the clinical data to obtain target data;
[0102] Performing data analysis on the target data through a trained deep learning model to obtain an analysis vector, wherein the deep learning model is a model trained based on an activation model, a pooling model, and a fully connected model connected in sequence;
[0103] Vector analysis is performed according to the analysis vector to obtain an analysis result.
[0104] The clinical data analysis method proposed in this embodiment obtains clinical data, wherein the clinical data includes biochemical indicators and / or electrocardiogram data of the target user, and then preprocesses the clinical data to obtain target data, and performs data analysis on the target data through a trained deep learning model to obtain an analysis vector, wherein the deep learning model is a model trained based on an activation model, a pooling model, and a fully connected model connected in sequence, and finally performs vector analysis based on the analysis vector to obtain an analysis result. The deep learning model trained based on the activation model, the pooling model, and the fully connected model can realize efficient and accurate data analysis of the clinical data, thereby obtaining an analysis result with improved analysis accuracy.
[0105] In one embodiment, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:
[0106] Acquiring clinical data, wherein the clinical data includes biochemical indicators and / or electrocardiogram data of the target user;
[0107] Preprocessing the clinical data to obtain target data;
[0108] Performing data analysis on the target data through a trained deep learning model to obtain an analysis vector, wherein the deep learning model is a model trained based on an activation model, a pooling model, and a fully connected model connected in sequence;
[0109] Vector analysis is performed according to the analysis vector to obtain an analysis result.
[0110] The clinical data analysis method proposed in this embodiment obtains clinical data, wherein the clinical data includes biochemical indicators and / or electrocardiogram data of the target user, and then preprocesses the clinical data to obtain target data, and performs data analysis on the target data through a trained deep learning model to obtain an analysis vector, wherein the deep learning model is a model trained based on an activation model, a pooling model, and a fully connected model connected in sequence, and finally performs vector analysis based on the analysis vector to obtain an analysis result. The deep learning model trained based on the activation model, the pooling model, and the fully connected model can realize efficient and accurate data analysis of the clinical data, thereby obtaining an analysis result with improved analysis accuracy.
[0111] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can refer to the relevant descriptions on the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.
[0112] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
[0113] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.
[0114] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.
Claims
1. A clinical data analysis method, characterized in that: The clinical data analysis method comprises: Acquiring clinical data, wherein the clinical data includes biochemical indicators and / or electrocardiogram data of the target user; Preprocessing the clinical data to obtain target data; Performing data analysis on the target data through a trained deep learning model to obtain an analysis vector, wherein the deep learning model is a model trained based on an activation model, a pooling model, and a fully connected model connected in sequence; Vector analysis is performed according to the analysis vector to obtain an analysis result.
2. The clinical data analysis method according to claim 1, characterized in that: The activation model includes a first convolutional layer, an activation layer, and a first pooling layer connected in sequence. The step of performing data analysis on the target data using the trained deep learning model to obtain an analysis vector includes: Inputting the target data into a first convolutional layer for convolution processing, and obtaining a first processing result output by the first convolutional layer; Inputting the first convolution result into the activation layer for activation processing, and inputting the activation processing result output by the activation layer into the first pooling layer for pooling processing to obtain a second processing result; An analysis vector is obtained according to the second processing result, the pooling model and the fully connected model.
3. The clinical data analysis method according to claim 2, characterized in that: The pooling model includes a first pooling module and a second pooling module connected in sequence, the first pooling module includes a second convolutional layer and a second pooling layer connected in sequence, and the step of obtaining an analysis vector according to the second processing result, the pooling model and the fully connected model includes: Inputting the second processing result into the second convolutional layer for convolution processing to obtain a third processing result output by the second convolutional layer; Inputting the third processing result into the second pooling layer for pooling processing to obtain a fourth processing result output by the second pooling layer; The analysis vector is obtained according to the fourth processing result, the second pooling module and the fully connected model.
4. The clinical data analysis method according to claim 3, characterized in that: The second pooling module includes a third convolutional layer, a fourth convolutional layer, a fifth convolutional layer and a third pooling layer connected in sequence, and the step of obtaining the analysis vector according to the fourth processing result, the second pooling module and the fully connected model includes: Inputting the fourth processing result into the third convolution layer for convolution processing to obtain a fifth processing result output by the third convolution layer; Inputting the fifth processing result into the fourth convolutional layer for convolution processing to obtain a sixth processing result output by the fourth convolutional layer; Inputting the sixth processing result into the fifth convolutional layer for convolution processing to obtain a seventh processing result output by the fifth convolutional layer; Inputting the seventh processing result into the third pooling layer for pooling processing to obtain an eighth processing result output by the third pooling layer; The analysis vector is obtained according to the eighth processing result and the fully connected model.
5. The clinical data analysis method according to claim 4, characterized in that: The fully connected model includes a first fully connected layer, a first Dropout layer, a second fully connected layer, and a second Dropout layer connected in sequence, and the step of obtaining the analysis vector according to the eighth processing result and the fully connected model includes: Inputting the eighth processing result into the first fully connected layer for connection, to obtain a first connection result output by the first fully connected layer; Inputting the first connection result into the first Dropout layer for anti-overfitting processing to obtain a target result; Inputting the target result into the second fully connected layer for connection, and obtaining a second connection result output by the second fully connected layer; The second connection result is input into the second Dropout layer for anti-overfitting processing to obtain the analysis vector.
6. The clinical data analysis method according to claim 1, characterized in that: The step of preprocessing the clinical data to obtain target data includes: The electrocardiogram data is denoised using discrete wavelet packet transform and sliding average method to obtain denoised data. The Pan-Tompkins algorithm and wavelet packet transform are used to locate the characteristic wave groups of the denoised data and obtain the target data.
7. The clinical data analysis method according to any one of claims 1 to 6, characterized in that: The step of performing vector analysis according to the analysis vector to obtain an analysis result comprises: By performing vector analysis on the analysis vector, a symptom portrait corresponding to the clinical data is generated; or, by performing vector analysis on the analysis vector, a symptom analysis text corresponding to the clinical data is generated.
8. A clinical data analysis device, characterized in that: The clinical data analysis device comprises: An acquisition module, used for acquiring clinical data, wherein the clinical data includes biochemical indicators and / or electrocardiogram data of the target user; A preprocessing module, used for preprocessing the clinical data to obtain target data; A vector module, used to perform data analysis on the target data through a trained deep learning model to obtain an analysis vector, wherein the deep learning model is a model trained based on an activation model, a pooling model, and a fully connected model connected in sequence; The analysis module is used to perform vector analysis according to the analysis vector to obtain an analysis result.
9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the clinical data analysis method according to any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the clinical data analysis method according to any one of claims 1 to 7 are implemented.