Internet-based postpartum recovery health assessment system
Through the Internet-based postpartum recovery health assessment system, the random forest regression model and health parameter change curve are used to realize real-time monitoring and feedback of maternal health status, solving the problem of lack of continuity and real-time evaluation in the existing technology, and providing Accurate evaluation criteria and timely health monitoring.
Patent Information
- Application Number
- CN202411062549.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-05
- Publication Date
- 2025-01-03
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing postpartum recovery health assessment methods lack continuity and real-timeness, rely on the patient's subjective recall and doctor's discrete observations, and cannot provide continuous and objective monitoring data, resulting in insufficient dynamic assessment of the maternal recovery process.
The postpartum recovery health assessment system based on the Internet is adopted to obtain user data through the data collection module, establish a theoretical prediction model based on random forest regression, predict postpartum health parameter data, and fit the health parameter change curve through the individual prediction module to analyze and evaluate the results. Provide real-time feedback.
Accurate, objective, continuous and real-time monitoring of maternal health status is achieved, problems and needs are discovered in a timely manner, and problems lack continuity and real-time nature of postpartum recovery health assessment in the existing technology are solved.
Smart Images
Figure CN119252472A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of postpartum recovery monitoring, and in particular to an Internet-based postpartum recovery health assessment system. Background Art
[0002] Postpartum recovery and health assessment is a comprehensive evaluation of the mother's physical condition after childbirth. The significance of postpartum recovery and health assessment is to help doctors and patients fully understand the progress and changes of postpartum recovery, identify problems and needs in a timely manner, and provide a basis for formulating personalized rehabilitation plans. Through the assessment, doctors can adjust treatment plans and rehabilitation guidance in a timely manner to improve the mother's quality of life and health level.
[0003] However, in the existing medical system, there are still some problems in the assessment of postpartum recovery. First, the existing assessment methods mainly rely on the patient's subjective recollection and the doctor's discrete observation, and cannot provide continuous and objective monitoring data. Secondly, the existing assessment methods usually only evaluate through data obtained during maternal physical examinations and regular follow-up, and cannot provide real-time monitoring and feedback. This method limits the comprehensive and dynamic assessment of the maternal recovery process, which may lead to delayed capture and processing of changes in the condition.
[0004] Therefore, there is an urgent need for a solution that can continuously and real-timely assess postpartum recovery health. Summary of the invention
[0005] In view of this, it is necessary to provide an Internet-based postpartum health recovery assessment system to solve the problem of lack of continuity and real-time performance in the assessment of postpartum health recovery in the prior art.
[0006] The present invention provides an Internet-based postpartum recovery health assessment system, comprising: A data collection module is used to obtain user data of all users, wherein the user data includes individual characteristic data, postpartum status data, and health parameter data at different time points; a model building module is used to build a theoretical prediction model based on random forest regression according to the user data of all users, wherein the theoretical prediction model is used to predict the health parameter data of the user at a target time point during the postpartum recovery period according to the individual characteristic data and postpartum status data; Theoretical prediction module, used to obtain user data of the target user, and based on the theoretical prediction model and the user data of the target user, obtain theoretical health parameter data of the target user at a target time point during the postpartum recovery period; An individual prediction module is used to obtain the actual health parameter data of the target user, and to fit the health parameter change prediction curve of the target user based on the actual health parameter data; The analysis and evaluation module is used to obtain a health evaluation result based on the difference between the health parameter change prediction curve and the theoretical health parameter data.
[0007] Furthermore, the method of establishing a theoretical prediction model based on random forest regression according to the user data of all users includes: According to the user data of all users, a sample is established based on the user data of each user; Dividing a training set and a test set according to a plurality of samples, and randomly extracting samples from the training set based on stratified sampling to obtain a plurality of sample sets; Based on Gaussian sampling, a decision tree is established for each sample set; Build a random forest model based on multiple decision trees; Based on the test set, the random forest model is verified and optimized to obtain the theoretical prediction model.
[0008] Furthermore, the user data also includes health markers that characterize whether the user is healthy after childbirth recovery; the training set and the test set are divided according to the multiple samples, and samples are randomly selected from the training set based on stratified sampling to obtain multiple sample sets, including: Based on the preset ratio, multiple samples are divided into training sets and test sets; According to the health sign, the training set is divided into healthy samples and abnormal samples; based on the Bootstrap method, samples of the same set are randomly selected from the healthy samples or the abnormal samples for multiple times to obtain multiple sample sets.
[0009] Furthermore, based on Gaussian sampling, a decision tree is established for each sample set, including: According to the target sample set, an input matrix is established, wherein each column of the input matrix corresponds to an individual characteristic data, a postpartum status data and a time point, and each row of the input matrix corresponds to a sample; According to the target sampling set, an output vector is established, wherein each element of the output vector corresponds to a health parameter data; Calculate the standard deviation of each column in the input matrix; Generate a set of random numbers for each column according to the standard deviation of each column, each set of random numbers obeys a Gaussian distribution with a mean of zero and a variance equal to the standard deviation of the corresponding column, and the number of each set of random numbers is the same as the number of rows of the input matrix; Adding each group of random numbers to the corresponding column of the input matrix to obtain an optimized input matrix; A decision tree corresponding to the target sample set is established according to the optimized input matrix and the output vector.
[0010] Furthermore, the processing of the random forest model includes: Taking individual characteristic data, postpartum status data and time point as input data, the input data is input into each decision tree for decision making, and an output result is obtained respectively; According to the constituent features of each decision tree, different weights are assigned to each output result; According to each output result and the corresponding weight, the output result of the random forest model is obtained.
[0011] Furthermore, the constituent feature is the height of the decision tree; and according to the constituent feature of each decision tree, different weights are assigned to each output result, including: The weight corresponding to each output result is obtained according to the following formula: $w_i = e^{-\alpha * h}$ in, $w_i$ For the i The weight corresponding to the output of each decision tree, α is the parameter that controls the decay rate, and h is the height of the i-th decision tree.
[0012] Furthermore, according to the weight corresponding to each output result, the output results of the random forest model are obtained, including: The output of the random forest model is obtained according to the following formula: $RF\_output = \sum(w_i * y_i)$ in, $RF\_output$ is the output result of the random forest model, $y_i$ For the i The output of a decision tree.
[0013] Furthermore, the individual characteristic data include height, initial weight and age; the postpartum status data include blood pressure, heart rate and body temperature; and the health parameter data include weight, uterine size, wound healing degree and breast fullness.
[0014] The beneficial effects of the present invention are: The present invention provides a postpartum recovery health assessment system based on the Internet, which can be used in a medical system, collects data of all users based on the Internet of a hospital, and establishes a theoretical prediction model based on the user data of all users. The theoretical prediction model can obtain health parameter data of a specific time node during the postpartum recovery period according to the individual characteristic data and postpartum status data of the user, and at the same time, each user can establish its actual health parameter change curve through the actual health parameter data, and then the system obtains the health assessment result according to the comparison between the health parameter change prediction curve and the predicted theoretical health parameter data, and gives relevant suggestions. Compared with the prior art, the present invention uses the theoretical prediction model established by the Internet to provide each user with an accurate and objective evaluation standard, and at the same time realizes the continuous and real-time monitoring of the patient (i.e., the user in the system) through the health parameter change prediction curve. The combination of the two realizes the real-time monitoring and feedback of postpartum recovery, and then timely discovers problems and needs, solving the problem of lack of continuity and real-time in the evaluation of postpartum recovery health in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] Figure 1 A system architecture diagram of an embodiment of an Internet-based postpartum recovery health assessment system provided by the present invention; Figure 2 A schematic diagram of a flow chart of an embodiment of an Internet-based postpartum recovery health assessment system provided by the present invention. DETAILED DESCRIPTION
[0016] 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.
[0017] Combination Figure 1 and Figure 2 As shown, a specific embodiment of the present invention discloses an Internet-based postpartum recovery health assessment system, which includes: The data collection module 110 is used to obtain user data of all users, wherein the user data includes individual characteristic data, postpartum status data, and health parameter data at different time points; A model building module 120 is used to build a theoretical prediction model based on random forest regression according to the user data of all users, wherein the theoretical prediction model is used to predict the health parameter data of the user at a target time point during the postpartum recovery period according to the individual characteristic data and the postpartum status data; Theoretical prediction module 130 is used to obtain user data of the target user, and obtain theoretical health parameter data of the target user at a target time point during the postpartum recovery period based on the theoretical prediction model and the user data of the target user; The individual prediction module 140 is used to obtain the actual health parameter data of the target user, and obtain the health parameter change prediction curve of the target user based on the nonlinear regression fitting according to the actual health parameter data; Analysis and evaluation module 150, used to predict the health parameter change curve and the theoretical The difference in health parameter data is used to obtain health assessment results.
[0018] Compared with the existing technology, the present invention uses the theoretical prediction model established by the Internet to provide each user with accurate and objective evaluation standards, and at the same time realizes continuous and real-time monitoring of patients (i.e., users in the system) through the health parameter change prediction curve. The combination of the two realizes real-time monitoring and feedback of postpartum recovery, thereby promptly discovering problems and needs, and solving the problem of lack of continuity and real-time performance in the evaluation of postpartum health recovery in the existing technology.
[0019] The system provided by the present invention can be run in the server of the medical system, and can be deployed in the computers in each consulting room of the hospital, the mobile phones of patients (users in the basic system), monitoring instruments and other terminals to obtain user data.
[0020] In the data collection module 110 of the present invention, individual characteristic data refers to parameters that can describe the individual characteristics of the user, such as height, initial weight, age, etc. Postpartum status data refers to data that the user needs to monitor after giving birth, such as blood pressure, heart rate, and body temperature. The health parameter data refers to data that can be used to evaluate the health level and can be graphically represented by a curve, such as weight, uterine size, wound healing degree, and breast fullness. The curve formed by the changes in these parameters is the health parameter change prediction curve. The health parameter change prediction curve can be used to more intuitively observe and analyze the change trend. The following is the role of the health parameter change curves corresponding to the four health parameters: Weight change prediction curve: By plotting the patient's weight over time, the speed and trend of weight recovery can be observed. A downward trend in the curve may indicate weight loss and return to pre-pregnancy levels.
[0021] Predicted uterine size change curve: Over time, plotting the change in uterine size as it recovers after childbirth can help assess the progress of uterine recovery. Normally, the uterus will gradually decrease in size until it returns to a non-pregnant state.
[0022] Wound healing prediction curve: For example, for surgical incisions during the postpartum period, a curve can be drawn by tracking the wound healing process. The shape and healing speed of the curve can be used to evaluate the degree of wound recovery.
[0023] Breast fullness prediction curve: For breastfeeding women, plotting the change in breast fullness over time can help assess breast recovery. An upward trend in the curve may indicate breast fullness and gradual recovery of breast function.
[0024] It is understandable that in addition to the above examples, individual characteristic data, postpartum status data, and health parameter data may also be other data according to actual needs.
[0025] In addition, the time points in the data collection module 110 are some key time points during the postpartum recovery period, and can also be some special time points designated by humans.
[0026] For example, the time of postpartum recovery varies due to individual differences and various factors, and can generally be divided into the following stages: Phase 1 (0-6 weeks): This phase is also known as the postpartum recovery period or puerperium. It usually takes 6 to 8 weeks for the uterus to return to a non-pregnant state. During this period, the uterus will gradually shrink and return to its original size, and postpartum bleeding will gradually decrease. The mother's body will also gradually recover, but she still needs to pay attention to rest, eat a proper diet, and avoid fatigue.
[0027] Phase 2 (after 6 weeks): This phase is the middle stage of postpartum recovery. Usually 8 to 12 weeks after delivery, many women will feel an increase in physical strength and emotional stability. At this stage, women can gradually resume their daily life and light physical activities. However, the specific recovery time varies from person to person, and some women may need longer.
[0028] Stage 3 (3-6 months): This stage is the late stage of postpartum recovery. At this stage, the uterus, pelvic floor muscles and ligaments of most mothers have basically recovered. The body gradually returns to its pre-pregnancy state, and you can start to gradually increase the intensity and time of exercise. However, for some mothers, it may take longer to fully recover.
[0029] Then, the above-mentioned time points can be the 6th week after delivery, the 3rd month after delivery, and the 6th month after delivery.
[0030] In a preferred embodiment, the model building module 120 establishes a theoretical prediction model based on random forest regression according to the user data of all users, specifically including: According to the user data of all users, a sample is established based on the user data of each user; Dividing a training set and a test set according to a plurality of samples, and randomly extracting samples from the training set based on stratified sampling to obtain a plurality of sample sets; Based on Gaussian sampling, a decision tree is established for each sample set; Build a random forest model based on multiple decision trees; Based on the test set, the random forest model is verified and optimized to obtain the theoretical prediction model.
[0031] Specifically, in a preferred embodiment, the user data also includes a health sign indicating whether the user is healthy after childbirth recovery; the training set and the test set are divided according to the multiple samples, and samples are randomly selected from the training set based on stratified sampling to obtain multiple sample sets, including: Based on the preset ratio, multiple samples are divided into training sets and test sets; According to the health sign, the training set is divided into healthy samples and abnormal samples; Based on the Bootstrap method, samples of the same set are randomly selected from the healthy samples or the abnormal samples for multiple times to obtain multiple sample sets.
[0032] In the above process, the preset ratio is the ratio of the number of samples in the training set and the test set, for example, 9:1. The health mark is used to indicate whether the user corresponding to the user data has recovered normally. Then the healthy samples and abnormal samples divided according to the health mark represent the data of healthy users recovered and the data of users with recovery problems, respectively. The decision tree established based on the combination of two or more samples can be more targeted and better able to identify whether the user is in a normal recovery state, making the theoretical prediction model finally established more accurate.
[0033] The Bootstrap method in this embodiment is the self-service sampling method (Bootstrap): for a training set of size N, N samples are randomly extracted from it with replacement to form a new sample set. The extracted samples will appear repeatedly, while some samples may not appear in the new sample set.
[0034] In a preferred embodiment, based on Gaussian sampling, a decision tree is established for each sample set, specifically including: According to the target sample set, an input matrix is established, wherein each column of the input matrix corresponds to an individual characteristic data, a postpartum status data and a time point, and each row of the input matrix corresponds to a sample; According to the target sampling set, an output vector is established, wherein each element of the output vector corresponds to a health parameter data; Calculate the standard deviation of each column in the input matrix; Generate a set of random numbers for each column according to the standard deviation of each column, each set of random numbers obeys a Gaussian distribution with a mean of zero and a variance equal to the standard deviation of the corresponding column, and the number of each set of random numbers is the same as the number of rows of the input matrix; Adding each group of random numbers to the corresponding column of the input matrix to obtain an optimized input matrix; A decision tree corresponding to the target sample set is established according to the optimized input matrix and the output vector.
[0035] The present invention also provides a preferred embodiment to illustrate the above process in more detail: Assume that the target sample set includes the data of the following three users: User 1: height 180cm, initial weight 70kg, age 25, blood pressure 120 / 80, heart rate 70 beats / minute, body temperature 36.5 degrees Celsius, postpartum time point 0 days, the corresponding weight health parameter is 70kg; User 2: Height 165cm, initial weight 55kg, age 30, blood pressure 110 / 70, heart rate 75 beats / minute, body temperature 36.8 degrees Celsius, 10 days postpartum, corresponding weight health parameters are 52kg; User 3: height 175cm, initial weight 65kg, age 35 years old, blood pressure 130 / 90, heart rate 80 beats / minute, body temperature 37.0 degrees Celsius, postpartum time point 30 days, the corresponding weight health parameter is 60kg; Then the input matrix can be established as follows (where blood pressure is selected as systolic pressure):
[0036] The corresponding output vector is: The standard deviation of each column in the input matrix is: Standard deviation of height: 6.454972243679028 Standard deviation of weight: 7.745966692414834 Standard deviation of age: 5 Standard deviation of blood pressure: 7.0710678118654755 Standard deviation of heart rate: 5 Standard deviation of body temperature: 0.25370814016737187 Standard deviation of postpartum time points: 14.1421356237 The random number matrix generated according to the standard deviation is:
[0037] The final optimized input matrix is:
[0038] It is understandable that the above data is only for the convenience of understanding and use, and there may be some inaccuracies. In practice, the data can be further processed by normalization according to specific needs. In addition, how to establish a decision tree corresponding to the target sample set based on the optimized input matrix and the output vector is an existing technology that can be understood by those skilled in the art, so it will not be explained in detail in this article.
[0039] The purpose of optimizing the input matrix based on Gaussian sampling in the present invention is to introduce a certain randomness into the sample, thereby increasing the diversity of the model. Gaussian sampling is usually used to process data sets with higher feature dimensions.
[0040] Furthermore, in a preferred embodiment, the processing of the random forest model includes: Taking individual characteristic data, postpartum status data and time point as input data, the input data is input into each decision tree for decision making, and an output result is obtained respectively; According to the constituent features of each decision tree, different weights are assigned to each output result; According to each output result and the corresponding weight, the output result of the random forest model is obtained.
[0041] In the above process, the constituent features of the tree include the number of leaf nodes, the height, depth, and width of the tree. The constituent features of a decision tree reflect the predictive power of the decision tree. For example, the higher the height of the tree, the more factors it considers in its judgment, and the higher the accuracy. Therefore, when analyzing with a random forest, more weight should be given to the output of a decision tree with a higher height, so that the final output of the random forest is more accurate.
[0042] In a preferred embodiment, the constituent feature is the height of the decision tree; and the method of assigning different weights to each output result according to the constituent feature of each decision tree includes: The weight corresponding to each output result is obtained according to the following formula: $w_i = e^{-\alpha * h}$ in, $w_i$ For the i The weight corresponding to the output of each decision tree, α is the parameter that controls the decay rate, and h is the height of the i-th decision tree.
[0043] It is understandable that other strategies may be used to assign weights in practice. For example, the weight of each output result may be directly the height value of the corresponding decision tree.
[0044] Further, in a preferred embodiment, obtaining the output result of the random forest model according to the weight corresponding to each output result includes: obtaining the output result of the random forest model according to the following formula: $RF\_output = \sum(w_i * y_i)$ in, $RF\_output$ is the output result of the random forest model, $y_i$ For the i The output of a decision tree.
[0045] In a preferred embodiment, in the theoretical prediction module 130 and the individual prediction module 140, the user data and actual health parameter data of the target user can be obtained through smart watches, mobile phones, weight scales, some home monitoring instruments, hospital review, etc. In addition, how to fit the discrete actual health parameter data into a continuous health parameter change prediction curve is a prior art and will not be described in detail in this article.
[0046] Similarly, in the analysis and evaluation module 150, how to obtain the health assessment result based on the difference between the health parameter change prediction curve and the theoretical health parameter data can also be flexibly designed or adopted in any existing way according to actual needs. The focus of this article is on the architectural background and principles of the entire system, rather than the patient's diagnostic method. Therefore, this article does not go into detail on how to conduct the evaluation.
[0047] The present invention provides a postpartum recovery health assessment system based on the Internet, which can be used in a medical system, collects data of all users based on the Internet of a hospital, and establishes a theoretical prediction model based on the user data of all users. The theoretical prediction model can obtain health parameter data of a specific time node during the postpartum recovery period according to the individual characteristic data and postpartum status data of the user, and at the same time, each user can establish its actual health parameter change curve through the actual health parameter data, and then the system obtains the health assessment result according to the comparison between the health parameter change prediction curve and the predicted theoretical health parameter data, and gives relevant suggestions. Compared with the prior art, the present invention uses the theoretical prediction model established by the Internet to provide each user with an accurate and objective evaluation standard, and at the same time realizes the continuous and real-time monitoring of the patient (i.e., the user in the system) through the health parameter change prediction curve. The combination of the two realizes the real-time monitoring and feedback of postpartum recovery, and then timely discovers problems and needs, solving the problem of lack of continuity and real-time in the evaluation of postpartum recovery health in the prior art.
[0048] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.
[0049] 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 postpartum recovery health assessment system based on the Internet, characterized in that: include: A data collection module, used to obtain user data of all users, the user data including individual characteristic data, postpartum status data and health parameter data at different time points; A model building module, used to build a theoretical prediction model based on random forest regression according to the user data of all users, wherein the theoretical prediction model is used to predict the health parameter data of the user at a target time point during the postpartum recovery period according to the individual characteristic data and the postpartum status data; Theoretical prediction module, used to obtain user data of the target user, and based on the theoretical prediction model and the user data of the target user, obtain theoretical health parameter data of the target user at a target time point during the postpartum recovery period; An individual prediction module is used to obtain the actual health parameter data of the target user, and to fit the health parameter change prediction curve of the target user based on the actual health parameter data; The analysis and evaluation module is used to obtain a health evaluation result based on the difference between the health parameter change prediction curve and the theoretical health parameter data.
2. The Internet-based postpartum recovery health assessment system according to claim 1, characterized in that: The method of establishing a theoretical prediction model based on random forest regression according to the user data of all users includes: According to the user data of all users, a sample is established based on the user data of each user; Dividing a training set and a test set according to a plurality of samples, and randomly extracting samples from the training set based on stratified sampling to obtain a plurality of sample sets; Based on Gaussian sampling, a decision tree is established for each sample set; Build a random forest model based on multiple decision trees; Based on the test set, the random forest model is verified and optimized to obtain the theoretical prediction model.
3. The Internet-based postpartum recovery health assessment system according to claim 2, characterized in that: The user data also includes a health indicator indicating whether the user is healthy after childbirth recovery; The method of dividing a training set and a test set according to a plurality of samples, and randomly extracting samples from the training set based on stratified sampling to obtain a plurality of sample sets includes: Based on the preset ratio, multiple samples are divided into training sets and test sets; According to the health sign, the training set is divided into healthy samples and abnormal samples; Based on the Bootstrap method, samples of the same set are randomly selected from the healthy samples or the abnormal samples for multiple times to obtain multiple sample sets.
4. The Internet-based postpartum recovery health assessment system according to claim 2, characterized in that: Based on Gaussian sampling, a decision tree is established for each sample set, including: According to the target sample set, an input matrix is established, wherein each column of the input matrix corresponds to an individual characteristic data, a postpartum status data and a time point, and each row of the input matrix corresponds to a sample; According to the target sampling set, an output vector is established, wherein each element of the output vector corresponds to a health parameter data; Calculate the standard deviation of each column in the input matrix; Generate a set of random numbers for each column according to the standard deviation of each column, each set of random numbers obeys a Gaussian distribution with a mean of zero and a variance equal to the standard deviation of the corresponding column, and the number of each set of random numbers is the same as the number of rows of the input matrix; Adding each group of random numbers to the corresponding column of the input matrix to obtain an optimized input matrix; A decision tree corresponding to the target sample set is established according to the optimized input matrix and the output vector.
5. The Internet-based postpartum recovery health assessment system according to claim 2, characterized in that: The processing of the random forest model includes: Taking individual characteristic data, postpartum status data and time point as input data, the input data is input into each decision tree for decision making, and an output result is obtained respectively; According to the constituent features of each decision tree, different weights are assigned to each output result; According to each output result and the corresponding weight, the output result of the random forest model is obtained.
6. The Internet-based postpartum recovery health assessment system according to claim 5, characterized in that: The constituent feature is the height of the decision tree; and according to the constituent features of each decision tree, different weights are assigned to each output result, including: The weight corresponding to each output result is obtained according to the following formula: w_i = exp(-α * h) in, w_i For the i The weight corresponding to the output of each decision tree, α is the parameter that controls the decay rate, and h is the height of the i-th decision tree.
7. The Internet-based postpartum recovery health assessment system according to claim 6, characterized in that: According to the weight corresponding to each output result, the output results of the random forest model are obtained, including: The output of the random forest model is obtained according to the following formula: RF_output = Σ(w_i * y_i) in, RF_output is the output result of the random forest model, y_i For the i The output of a decision tree.
8. The Internet-based postpartum recovery health assessment system according to claim 1, characterized in that: The individual characteristic data include height, initial weight and age; the postpartum status data include blood pressure, heart rate and body temperature; and the health parameter data include weight, uterine size, wound healing degree and breast fullness.