Disease prognosis risk early warning method and system based on data acquisition
Through support vector regression model and genetic algorithm based on data acquisition, combined with database systems and machine learning, the problem of incomplete and accurate disease prognosis risk assessment in the existing technology is solved, and more efficient and accurate disease prognosis risk warning is achieved. Ensure the security of patient data.
Patent Information
- Application Number
- CN202411415539.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-11
- Publication Date
- 2024-12-31
AI Technical Summary
In the prior art, the risk assessment of disease prognosis risk usually independently judges various factors, and does not consider the correlation between factors, resulting in the incomplete and accurate assessment.
Using a data acquisition-based method, combining support vector regression model and genetic algorithm, a database system is established, patient information is collected, and patient information is processed and stored, and the prognosis risk probability is predicted through machine learning models, and the risk level is feedback in real time for adjustment and intervention to ensure Data encryption is processed to protect privacy.
Achieve a more scientific and efficient early warning of disease prognosis risk, and evaluate it by linking multiple factors, improving prediction accuracy and safety, reducing dependence on medical staff experience, saving labor costs, and ensuring the safety of patient data .
Smart Images
Figure CN119230115A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of disease risk early warning, and more specifically to a disease prognosis risk early warning method and system based on data collection. Background Art
[0002] Disease prognosis is the understanding of a certain disease. In addition to understanding its clinical manifestations, laboratory tests and imaging, etiology, pathology and disease patterns, it is important to evaluate the short-term and long-term efficacy, recovery or progression of the disease based on the timing and method of treatment combined with new conditions discovered during the treatment operation. Disease prognosis is related to many factors such as the timing of treatment, the severity of the disease, medical level, comorbidities, the doctor's personal ability, constitution, age, whether the patient faces the disease or has cognitive ability to the disease, and whether to continue treatment. Determining the risk of disease prognosis is an important task in the medical field. However, the assessment of disease prognosis risk in the prior art is usually based on independent factors, without considering the relationship between factors.
[0003] Therefore, how to provide a disease prognosis risk warning method and system based on data collection, introduce advanced data analysis technology and database system, and more comprehensively and accurately evaluate the prognosis risk of the disease is an urgent problem that technicians in this field need to solve. Summary of the invention
[0004] In view of this, the present invention provides a disease prognosis risk warning method and system based on data collection, which combines data collection, advanced machine learning models and real-time feedback mechanisms to make disease prognosis risk warning more scientific and efficient.
[0005] In order to achieve the above object, the present invention adopts the following technical solution: a disease prognosis risk early warning method based on data collection, comprising:
[0006] Establish a database system to collect patient information, and process and store the collected patient information;
[0007] Inputting patient information into a trained prognostic risk analysis model to predict the patient's disease prognostic risk probability and achieve safety monitoring and early warning of the patient; the prognostic risk analysis model adopts a support vector regression model, and when establishing the support vector regression model, the hyperparameter optimization of the support vector regression model is performed through a genetic algorithm;
[0008] The risk level is determined according to the analysis results of the prognostic risk analysis model, and warning information is automatically generated according to the risk level, and the warning information is fed back to the patient and medical staff in real time, so as to make corresponding adjustments and interventions.
[0009] Preferably, it also includes: establishing a data encryption mechanism to ensure that all patient data is encrypted during the collection, storage and transmission process to protect patient privacy and data security.
[0010] Preferably, a database system is established to collect patient information, and the collected patient information is processed and stored, including:
[0011] Classifying the disease types in the patient information and establishing links between the same diseases;
[0012] According to the preset time step, the clinical data of existing patients are collected periodically to obtain the actively collected clinical data of patients;
[0013] Extract information from actively collected patient clinical data to generate supplementary patient data, which are then entered into the database system to generate patient disease update reports.
[0014] Preferably, the performing hyperparameter optimization on the support vector regression model by using a genetic algorithm comprises:
[0015] (1) Using the hyperparameters of the support vector regression model as elements to form a hyperparameter vector, randomly generating N different hyperparameter vectors as N different individuals in the genetic algorithm, the hyperparameters in the hyperparameter vector are the chromosomes of the individuals, encoding the chromosomes to obtain the genotype of the chromosomes, wherein N is an integer greater than 0;
[0016] (2) Setting the fitness function based on the patient information and its corresponding risk probability;
[0017] (3) Set the maximum number of iterations;
[0018] (4) calculating the fitness of the individual according to the fitness function, and selecting a number of individuals as parent individuals according to the fitness;
[0019] (5) performing a crossover operation on the parent individuals with a predetermined crossover probability pm, wherein the crossover operation refers to exchanging some genes of the corresponding chromosomes of the two parent individuals;
[0020] (6) performing a mutation operation on the parent individual after the crossover operation with a predetermined mutation probability pc, wherein the mutation operation refers to randomly transforming part of the genes of the parent individual;
[0021] Repeat steps (4) to (6) above until the maximum number of iterations is reached.
[0022] Preferably, the hyperparameters include a penalty coefficient C, an insensitive loss coefficient epsilon, and a width coefficient gamma of the support vector regression model.
[0023] Preferably, the fitness function is the mean square error between the predicted value and the actual value of the risk probability obtained by inputting the patient information into the support vector regression model corresponding to the hyperparameter vector, and the selection of several individuals as parent individuals based on the fitness adopts a roulette wheel selection strategy.
[0024] Using machine learning algorithms, with patient information as training data and the disease risk probability corresponding to the patient information as a marker column, a machine learning prediction model is established for training, and the relationship between patient information and disease risk probability is explored, so that the patient information can be input into the trained prognosis risk analysis model to obtain the predicted value of the disease risk probability, thereby avoiding reliance on the experience of medical staff, greatly saving manpower costs, and making the prediction of disease prognosis risk more accurate.
[0025] Preferably, when the predicted value of the risk probability reaches a preset threshold, the patient information data within the current preset time period is read and a warning message is sent to the patient and medical staff.
[0026] Preferably, the risk levels are divided into the following three categories:
[0027] Low risk: The patient's health condition is relatively stable and the risk probability is low;
[0028] Medium risk: The patient has certain health risks and needs further observation or some intervention measures;
[0029] High Risk: The patient is at high risk and faces a serious health problem or emergency, requiring immediate action.
[0030] Preferably, a disease prognosis risk early warning system based on data collection includes:
[0031] The data processing module is used to establish a database system, collect patient information, and process and store the collected patient information;
[0032] A data analysis module, used to input patient information into a trained prognostic risk analysis model, wherein the prognostic risk analysis model adopts a support vector regression model, and when establishing the support vector regression model, a genetic algorithm is used to perform hyperparameter optimization on the support vector regression model;
[0033] The risk warning module is used to determine the risk level according to the analysis results of the prognostic risk analysis model, automatically generate warning information according to the risk level, and feed back the warning information to patients and medical staff in real time.
[0034] It can be known from the above technical scheme that compared with the prior art, the present invention discloses a disease prognosis risk warning method and system based on data collection, including: establishing a database system, collecting patient information, and processing and storing the collected patient information; inputting patient information into a trained prognosis risk analysis model to predict the patient's disease prognosis risk probability, and realizing safety monitoring and early warning of the patient; the prognosis risk analysis model adopts a support vector regression model, and when establishing the support vector regression model, the support vector regression model is optimized by a genetic algorithm for hyperparameters; the risk level is determined according to the results of the prognosis risk analysis model, and early warning information is automatically generated according to the risk level, and the early warning information is fed back to the patient and medical staff in real time, so as to make corresponding adjustments and interventions. The present invention combines data collection, advanced machine learning models and real-time feedback mechanisms to make disease prognosis risk warnings more scientific and efficient. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] 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 embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.
[0036] Figure 1 A schematic flow chart of a disease prognosis risk warning method based on data collection provided by the present invention.
[0037] Figure 2 A schematic diagram of the structure of a disease prognosis risk warning system based on data collection provided by the present invention. DETAILED DESCRIPTION
[0038] 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 only 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.
[0039] The embodiment of the present invention discloses a disease prognosis risk early warning method based on data collection, such as Figure 1 As shown, including:
[0040] Establishing a database system to collect patient information, including basic information, clinical data and lifestyle data of the patient, and processing and storing the collected patient information;
[0041] Inputting patient information into a trained prognostic risk analysis model to predict the patient's disease prognostic risk probability and achieve safety monitoring and early warning of the patient; the prognostic risk analysis model adopts a support vector regression model, and when establishing the support vector regression model, the hyperparameter optimization of the support vector regression model is performed through a genetic algorithm;
[0042] The risk level (such as low, medium or high risk) is determined according to the analysis results of the prognostic risk analysis model, and early warning information is automatically generated according to the risk level, and the early warning information is fed back to the patient and medical staff in real time, so as to make corresponding adjustments and interventions.
[0043] At the same time, an intuitive user interface is designed to facilitate medical staff and patients to use and view risk analysis results;
[0044] In the embodiment of the present invention, (1) Establishment of a database system: collect relevant information of patients, including basic medical history, laboratory test results, imaging data, etc., to ensure the comprehensiveness and accuracy of the data. The collected patient information is standardized and stored for subsequent analysis. (2) Application of a prognostic risk analysis model: input the processed patient information into a trained prognostic risk analysis model, which uses a support vector regression (SVR) model to predict risk probability. When constructing a support vector regression model, a genetic algorithm is used to optimize the hyperparameters of the model to improve the prediction accuracy. (3) Determination of risk level and generation of early warning information: according to the output of the prognostic risk analysis model, the risk level of the patient is analyzed and can be divided into three different levels: low risk, medium risk and high risk. The system automatically generates early warning information based on the risk level to ensure that it can be fed back to patients and medical staff in a timely and accurate manner. (4) Real-time monitoring and intervention: realize safety monitoring of patients, timely detect risk changes, and provide relevant early warning information so that medical staff can respond quickly. Through real-time feedback, medical staff can take corresponding adjustment and intervention measures according to the specific situation of the patient to improve the quality of patient care.
[0045] Specifically, it also includes: establishing a data encryption mechanism to ensure that all patient data is encrypted during the collection, storage and transmission process.
[0046] (1) Install a patient data encryption system on the server side and set up a data verification service and a data management service respectively; the data verification service is used to verify the integrity and accuracy of the data, and the data management service is used to control the encryption and storage of patient data.
[0047] (2) Timestamp the original data and keep it intact and unchanged to ensure the time when the record was generated; then start the data verification service, control the encryption system to receive patient information, and verify the received patient information to form a new data value; when the verification is completed, generate a data fingerprint of the original patient information file; such as using MD5 or SHA-256 to generate a hash value.
[0048] (3) Start the data management service and control the encryption system to encrypt the eligible patient information according to the corresponding disease prognosis development. Encrypt the patient information every time it is read and the new patient information data is stored and returned, and use the regular verification method to verify the patient information; ensure the correctness of the data format and content.
[0049] (4) Perform corresponding encryption and storage operations on patient information.
[0050] Specifically, step 4.1: perform regular expression verification on the first element in the patient information (for example, name, ID number, etc.), encrypt the first element that passes the verification, generate first patient ciphertext data after encryption, and store the generated first patient ciphertext data in a targeted field in the patient information to ensure that the storage method of the ciphertext data is convenient for subsequent reading and decryption.
[0051] Step 4.2: Perform regular expression verification on the second element in the patient information (for example, contact information, address, etc.), encrypt the verified second element, generate second patient ciphertext data, and store it together with the first patient ciphertext data generated in step 4.1; create or update a storage structure that contains the first patient ciphertext data and the second patient ciphertext data to ensure data relevance.
[0052] Step 4.3: Repeat the steps 4.1 and 4.2 to encrypt and store the remaining elements of the patient information. Perform regular verification, encryption, and storage on each element in turn, and ensure that all patient ciphertext data can be associated and retrieved as needed each time it is stored. After all elements have been processed, perform an overall integrity and consistency check to ensure that all stored ciphertext data is available and that no data is lost or damaged.
[0053] (5) Read the encrypted and stored patient information, perform multiple decryptions on the encrypted fixed disease information, obtain the fixed disease information and verify the MD5 value and timestamp of the data to ensure the integrity and credibility of the data. After verification, read and use it as new data.
[0054] The data fingerprint generated in step (2) identifies the corresponding original patient information file, and stores the storage address of the original patient information file, the corresponding patient information content, and the generated data fingerprint.
[0055] In step (3), the data management service classifies, encrypts and stores different elements of the same disease information according to the elements of disease storage, ensuring that private data is only accessed when necessary.
[0056] The data encryption protection method provided in the embodiment of the present invention applies multiple asymmetric encryption technologies to the patient's disease prognosis risk warning work, and undergoes multiple encryptions when transmitting important patient information, which effectively improves and consolidates the security of the patient's personal information and protects the patient's privacy and data security.
[0057] Specifically, a database system is established to collect patient information, and the collected patient information is processed and stored, including:
[0058] Classifying the disease types in the patient information and establishing links between the same diseases;
[0059] Among them, the existing patient information is extracted to obtain the patient's basic information, clinical data and lifestyle data;
[0060] Generate corresponding disease labels based on the patient's basic information, clinical data, and lifestyle data;
[0061] Disease types are classified and stored according to disease labels, and indexes are created according to disease labels.
[0062] According to the preset time step, the clinical data of existing patients are collected periodically to obtain the actively collected clinical data of patients;
[0063] Among them, the time interval for data collection (such as weekly, monthly, etc.) is set according to clinical needs, and a scheduled task (such as Cron job) is used to automatically trigger data collection;
[0064] Actively collect patients' clinical data through questionnaires, mobile applications or online platforms, including information such as symptom changes and treatment responses. Or connect with the hospital's information system (such as an electronic medical record system) to obtain patients' clinical data periodically according to preset time.
[0065] Extract information from actively collected patient clinical data to generate supplementary patient data, which are then entered into the database system to generate patient disease update reports.
[0066] Among them, information extraction is performed on the actively collected clinical data of patients, the latest clinical data and lifestyle data contained therein are extracted, and supplementary patient data is generated accordingly;
[0067] According to the time sequence, the supplementary patient data is stored to obtain the disease evolution situation;
[0068] Compare the differences between the supplementary patient data and the existing patient information and generate a disease update report for the patient.
[0069] By classifying the patient information stored in the database system according to the disease type and periodically storing relevant data based on the subsequent development and evolution of the patient's disease, a full-cycle record of the disease type is achieved, and subsequent data search is convenient.
[0070] Specifically, the performing hyperparameter optimization on the support vector regression model by using a genetic algorithm includes:
[0071] (1) Using the hyperparameters of the support vector regression model as elements to form a hyperparameter vector, randomly generating N different hyperparameter vectors as N different individuals in the genetic algorithm, the hyperparameters in the hyperparameter vector are the chromosomes of the individuals, encoding the chromosomes to obtain the genotype of the chromosomes, wherein N is an integer greater than 0;
[0072] In the embodiment of the present invention, the penalty coefficient C, the insensitive loss coefficient epsilon and the width coefficient gamma form a hyperparameter vector [C, epsilon, gamma], take different values of C, epsilon, gamma, and randomly generate N different vectors [C, epsilon, gamma] as N different individuals, each of which is a chromosome of the individual. The chromosome can be encoded by various encoding methods of the prior art. Taking binary encoding as an example, the chromosome is encoded by converting the hyperparameters C, epsilon, and gamma in the vector [C, epsilon, gamma] into binary numbers respectively, and the numbers 0 or 1 in the binary numbers are the genes of the chromosome.
[0073] (2) Setting the fitness function based on the patient information and its corresponding risk probability;
[0074] The fitness function is set to the mean square error between the predicted value and the actual value of the disease risk probability obtained by inputting the patient information into the SVR model corresponding to the hyperparameter vector. The smaller the mean square error, the more accurate the prediction result of the SVR model formed by the hyperparameter vector, and the higher the fitness of the individual.
[0075] (3) Set the maximum number of iterations;
[0076] (4) calculating the fitness of the individual according to the fitness function, and selecting a number of individuals as parent individuals according to the fitness;
[0077] The embodiment of the present invention adopts a roulette selection strategy to select individuals. In the roulette selection strategy, the probability of an individual being selected is equal to the ratio of its fitness value to the sum of all individual fitness values. In a specific implementation, a random number in the interval [0, 1] can be randomly generated. If the random number is less than or equal to the probability of an individual being selected, the individual is selected.
[0078] (5) performing a crossover operation on the parent individuals with a predetermined crossover probability pm, wherein the crossover operation refers to exchanging some genes of the corresponding chromosomes of the two parent individuals;
[0079] The embodiment of the present invention generates a random number in the interval [0, 1], and performs a crossover operation when the random number is less than the crossover probability pm, and the crossover probability pm can be set to 0.9;
[0080] In the crossover operation, the parent individuals are first paired, and then some genes of the corresponding chromosomes of the two paired parent individuals are exchanged. This process is similar to the crossing of chromosomes in biology. When binary coding is used, the crossover operation is to exchange the binary digital strings on the two chromosomes. The crossover operator used in the crossover operation can be various crossover operators in the prior art.
[0081] (6) performing a mutation operation on the parent individual after the crossover operation with a predetermined mutation probability pc, wherein the mutation operation refers to randomly transforming part of the genes of the parent individual;
[0082] The embodiment of the present invention generates a random number in the interval [0, 1], and performs a mutation operation when the random number is less than the mutation probability pc, and the mutation probability pc is set to 0.1. The mutation operation refers to the random transformation of part of the genes of the parent individual, for example, when binary coding is used, it is to randomly change a gene from 0 to 1 or from 1 to 0. The mutation operator used in the mutation operation can be various mutation operators in the prior art.
[0083] Repeat steps (4) to (6) above until the maximum number of iterations is reached.
[0084] Determine whether the number of iterations has reached the set maximum number of iterations. If not, continue to execute the above processes (4) to (6). If it has reached, end the algorithm process.
[0085] Finally, the SVR model is constructed by using the hyperparameters found by the genetic algorithm, and the model is trained to obtain the optimal model. When constructing a machine learning prediction model, the embodiment of the present invention uses a genetic algorithm to find the optimal hyperparameters of the model, which solves the problem that the hyperparameters of the machine learning prediction model are difficult to select and improves the performance of the model.
[0086] Specifically, the hyperparameters include a penalty coefficient C, an insensitive loss coefficient epsilon, and a width coefficient gamma of the support vector regression model.
[0087] Specifically, the fitness function is the mean square error between the predicted value and the actual value of the risk probability obtained by inputting the patient information into the support vector regression model corresponding to the hyperparameter vector, and the selection of several individuals as parent individuals based on the fitness adopts a roulette wheel selection strategy.
[0088] Using machine learning algorithms, with patient information as training data and the disease risk probability corresponding to the patient information as a marker column, a machine learning prediction model is established for training, and the relationship between patient information and disease risk probability is explored, so that the patient information can be input into the trained prognosis risk analysis model to obtain the predicted value of the disease risk probability, thereby avoiding reliance on the experience of medical staff, greatly saving labor costs, and making the prediction of disease prognosis risk more accurate.
[0089] Specifically, when the predicted value of the risk probability reaches a preset threshold, the patient information data within the current preset time period is read and a warning message is sent to the patient and medical staff.
[0090] Specifically, the risk levels are divided into the following three categories:
[0091] Low Risk (Level 1): The patient's health condition is relatively stable and the risk probability is low; usually no immediate intervention is required.
[0092] Treatment measures: Regular monitoring and regular collection of patient information. Medical staff can provide corresponding guidance and suggestions during routine examinations.
[0093] Moderate Risk (Level 2): The patient has a certain health risk and needs further observation or some intervention measures; but it is not an emergency.
[0094] Treatment measures: Monitoring should be strengthened, regular reviews or adjustments to treatment plans should be conducted, medical staff should inform patients of relevant risks and precautions, and consider developing personalized management plans.
[0095] High Risk (Level 3): The patient is at high risk and faces a serious health problem or emergency, requiring immediate action.
[0096] Treatment measures: Emergency treatment, medical staff should quickly assess the patient's information status and take appropriate treatment measures, and perform emergency intervention when necessary. At the same time, provide clear guidance and information to patients and their families to ensure that they understand the current health risks and necessary response measures.
[0097] By classifying risk levels into these three categories, medical teams can manage patient risks more effectively and take appropriate measures quickly to ensure patient health and safety.
[0098] Specifically, the embodiment of the present invention also discloses a disease prognosis risk warning system based on data collection, such as Figure 2 As shown, including:
[0099] The data processing module is used to establish a database system, collect patient information, and process and store the collected patient information;
[0100] A data analysis module, used to input patient information into a trained prognostic risk analysis model, wherein the prognostic risk analysis model adopts a support vector regression model, and when establishing the support vector regression model, a genetic algorithm is used to perform hyperparameter optimization on the support vector regression model;
[0101] The risk warning module is used to determine the risk level according to the analysis results of the prognostic risk analysis model, automatically generate warning information according to the risk level, and feed back the warning information to patients and medical staff in real time.
[0102] It also includes: a data encryption module, which is used to establish a data encryption mechanism to ensure that all patient data is encrypted during the collection, storage and transmission process.
[0103] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.
[0104] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may 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 the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A disease prognosis risk early warning method based on data collection, characterized in that: include: Establish a database system to collect patient information, and process and store the collected patient information; Inputting patient information into a trained prognostic risk analysis model, wherein the prognostic risk analysis model adopts a support vector regression model, and when establishing the support vector regression model, performing hyperparameter optimization on the support vector regression model by using a genetic algorithm; The risk level is determined according to the analysis results of the prognostic risk analysis model, and early warning information is automatically generated according to the risk level, and the early warning information is fed back to the patient and medical staff in real time.
2. A disease prognosis risk early warning method based on data collection according to claim 1, characterized in that: Also includes: Establish a data encryption mechanism to ensure that all patient data is encrypted during the collection, storage and transmission process.
3. The disease prognosis risk early warning method based on data collection according to claim 1, characterized in that: Establish a database system to collect patient information, and process and store the collected patient information, including: Classifying the disease types in the patient information and establishing links between the same diseases; According to the preset time step, the clinical data of existing patients are collected periodically to obtain the actively collected clinical data of patients; Extract information from actively collected patient clinical data to generate supplementary patient data, which are then entered into the database system to generate patient disease update reports.
4. The disease prognosis risk early warning method based on data collection according to claim 3, characterized in that: The method of optimizing the hyperparameters of the support vector regression model by using a genetic algorithm includes: (1) Using the hyperparameters of the support vector regression model as elements to form a hyperparameter vector, randomly generating N different hyperparameter vectors as N different individuals in the genetic algorithm, the hyperparameters in the hyperparameter vector are the chromosomes of the individuals, encoding the chromosomes to obtain the genotype of the chromosomes, wherein N is an integer greater than 0; (2) Setting the fitness function based on the patient information and its corresponding risk probability; (3) Set the maximum number of iterations; (4) calculating the fitness of the individual according to the fitness function, and selecting a number of individuals as parent individuals according to the fitness; (5) performing a crossover operation on the parent individuals with a predetermined crossover probability pm, wherein the crossover operation refers to exchanging some genes of the corresponding chromosomes of the two parent individuals; (6) performing a mutation operation on the parent individual after the crossover operation with a predetermined mutation probability pc, wherein the mutation operation refers to randomly transforming part of the genes of the parent individual; Repeat steps (4) to (6) above until the maximum number of iterations is reached.
5. The disease prognosis risk early warning method based on data collection according to claim 1, characterized in that: The hyperparameters include a penalty coefficient C, an insensitive loss coefficient epsilon, and a width coefficient gamma of the support vector regression model.
6. The disease prognosis risk early warning method based on data collection according to claim 4, characterized in that: The fitness function is the mean square error between the predicted value and the actual value of the risk probability obtained by inputting the patient information into the support vector regression model corresponding to the hyperparameter vector, and the roulette wheel selection strategy is adopted for selecting several individuals as parent individuals according to the fitness.
7. A disease prognosis risk early warning method based on data collection according to claim 6, characterized in that: When the predicted value of the risk probability reaches the preset threshold, the patient information data within the current preset time period is read and an early warning message is sent to the patient and medical staff.
8. The disease prognosis risk early warning method based on data collection according to claim 1, characterized in that: The risk levels are divided into the following three categories: Low risk: The patient's health condition is relatively stable and the risk probability is low; Medium risk: The patient has certain health risks and needs further observation or some intervention measures; High Risk: The patient is at high risk and faces a serious health problem or emergency, requiring immediate action.
9. A disease prognosis risk early warning system based on data collection, characterized in that: A disease prognosis risk early warning method based on data collection according to any one of claims 1 to 8 is applied, comprising: The data processing module is used to establish a database system, collect patient information, and process and store the collected patient information; A data analysis module, used to input patient information into a trained prognostic risk analysis model, wherein the prognostic risk analysis model adopts a support vector regression model, and when establishing the support vector regression model, a genetic algorithm is used to perform hyperparameter optimization on the support vector regression model; The risk warning module is used to determine the risk level according to the analysis results of the prognostic risk analysis model, automatically generate warning information according to the risk level, and feed back the warning information to patients and medical staff in real time.