Vaginal delivery complication risk prediction model construction method and system, electronic equipment and storage medium

Abstract

The invention discloses a vaginal delivery complication risk prediction model construction method and system, electronic equipment and a storage medium. The prediction model construction method comprises the following steps: step S1, obtaining feature data of pregnant and lying-in women; s2, acquiring risk indexes of the characteristic data of the pregnant and lying-in women; s3, dividing the set pregnant and lying-in woman characteristic data into a training data set and a test data set; and S4, modeling the divided data set in a random forest model manner by using the data set obtained in the step S3. According to the vaginal delivery complication risk prediction model construction method and system, the electronic equipment and the storage medium, occurrence conditions of various vaginal delivery common complications can be more comprehensively predicted, and complication occurrence risk prediction and evaluation can be objectively, rapidly and effectively performed according to different pregnant and lying-in woman characteristic data.

Description

Technical Field

[0001] The present invention belongs to the technical field of clinical prediction model construction, and relates to a prediction model construction method, and in particular to a method, system, electronic device and storage medium for constructing a vaginal delivery complication risk prediction model. Background Art

[0002] Clinical prediction models are quantitative tools for risk and benefit assessment, among which prognostic models are widely used to predict disease complications. Complication risk prediction models quantitatively assess the risk of complications in patients during a certain disease or treatment by analyzing complication-related risk factor data. At present, research on complication prediction models at home and abroad mainly focuses on the fields of diabetic complications and postoperative complications of surgery, such as complications after pelvic hip fracture surgery, complications after thoracoscopic radical resection of lung cancer, and complications after colorectal cancer surgery.

[0003] At present, there is no comprehensive prediction model for the risk of common complications of vaginal delivery, which makes it difficult to achieve the important clinical value of early identification and intervention of common complications of vaginal delivery. The risk prediction models related to common complications of vaginal delivery are mainly aimed at single complications, such as postpartum hemorrhage risk prediction model and obstetric laceration risk prediction model. There are many assessment tools for the risk of various single complications of vaginal delivery, and the assessment quality and effect are uneven, and the clinical utilization rate is low. The reason is that the construction of vaginal delivery complication risk prediction models is mostly based on retrospective cohort studies and some research samples are limited to a certain specific population. There is a lack of strict internal and external validation process, and the model prediction performance needs to be further improved.

[0004] In view of this, there is an urgent need to design a new method for constructing clinical prediction models in order to overcome at least some of the above-mentioned defects of existing methods for constructing clinical prediction models. Summary of the invention

[0005] The present invention provides a method, system, electronic device and storage medium for constructing a vaginal delivery complication risk prediction model, which can more comprehensively predict the occurrence of multiple common vaginal delivery complications and objectively, quickly and effectively make a risk prediction assessment of the occurrence of complications based on different maternal and child characteristic data.

[0006] To solve the above technical problem, according to one aspect of the present invention, the following technical solution is adopted:

[0007] A method for constructing a risk prediction model for vaginal delivery complications, the method comprising:

[0008] Step S1, obtaining maternal characteristic data;

[0009] Step S2, obtaining risk indicators of maternal and infant characteristic data;

[0010] Step S3, dividing the set maternal and infant characteristic data into a training data set and a test data set for modeling;

[0011] Step S4: Using the data set obtained in step S3, the divided data set is modeled in the manner of a random forest model.

[0012] As an implementation mode of the present invention, step S4 includes:

[0013] The sample selection step is to select sample data from the set database;

[0014] a data preprocessing step, preprocessing the sample data selected by the sample selection unit;

[0015] Feature selection step, selecting feature data; the selected feature data include antenatal data (general condition of pregnant women, previous history of obstetrics and gynecology, and current pregnancy), delivery data (information of this delivery process, newborn conditions, and maternal conditions during delivery), and postpartum data (occurrence of common complications of vaginal delivery);

[0016] Variable selection step: select variable data; use Logistic regression and Lasso regression respectively to obtain the variables that have the most significant impact on the prediction results, and finally determine the best combination of predictive variables;

[0017] Modeling steps: establishing mathematical models; conducting mathematical modeling based on the best combination of predictive variables;

[0018] In the hyperparameter tuning step, a grid parameter search method is used to iterate through all possible parameter combinations to perform parameter combination tuning.

[0019] The variable importance ranking step uses the average prediction accuracy reduction value and the average Gini index reduction value to rank the importance of the variables in the random forest model;

[0020] The model prediction efficiency evaluation step evaluates the overall prediction performance of the model through the model's prediction accuracy, sensitivity, and specificity; and evaluates the model's prediction effect from three aspects: the model's discrimination, calibration, and clinical practicality.

[0021] As an implementation of the present invention, in step S4, hyperparameter tuning is performed, and the parameter range for random forest modeling is assumed to be: number of decision trees (ntree) = c (100, 200, 300, 400, 500); number of node partition features (mtry) = c (2, 3, 4, 5). Finally, the optimal model parameter configuration is obtained as follows: number of decision trees (ntree) = 200, number of node partition features (mtry) = 4, and the trained model has the best prediction performance.

[0022] As an embodiment of the present invention, the method further comprises:

[0023] Step S5, performing variable importance analysis on the data set obtained in step S3;

[0024] The importance of variables in the random forest model was ranked using the average prediction accuracy reduction value and the average Gini index reduction value;

[0025] Step S6, evaluating the prediction performance of the model;

[0026] The overall prediction performance of the model was evaluated through its prediction accuracy, sensitivity and specificity. The prediction effect of the model was evaluated from three aspects: the discrimination, calibration and clinical practicality of the model.

[0027] According to another aspect of the present invention, the following technical solution is adopted: a system for constructing a risk prediction model for vaginal delivery complications, the system comprising:

[0028] A characteristic data acquisition module is used to obtain characteristic data of pregnant women;

[0029] A risk indicator acquisition module is used to obtain risk indicators of maternal and infant characteristic data;

[0030] A data set allocation module is used to divide the set maternal and infant characteristic data into a training data set and a test data set for modeling;

[0031] The prediction model building module is used to model the divided data set in the form of a random forest model to build a vaginal delivery complication risk prediction model.

[0032] As an implementation mode of the present invention, the prediction model building module includes:

[0033] A sample selection unit, used to select sample data from a set database;

[0034] A data preprocessing unit, used for preprocessing the sample data selected by the sample selection unit;

[0035] The feature selection unit is used to select feature data; the selected feature data include antenatal data (general condition of pregnant women, previous history of obstetrics and gynecology, and current pregnancy), delivery data (information of the current delivery process, condition of the newborn, and condition of the mother during delivery), and postpartum data (occurrence of common complications of vaginal delivery);

[0036] The variable selection unit is used to select variable data; Logistic regression and Lasso regression are used to obtain the variables that have the most significant impact on the prediction results, and finally determine the best combination of predictive variables;

[0037] A modeling unit is used to establish a mathematical model and perform mathematical modeling according to the best combination of predictive variables;

[0038] The hyperparameter tuning unit is used to use a grid parameter search method to iterate through all possible parameter combinations to perform parameter combination tuning.

[0039] As an embodiment of the present invention, the vaginal delivery complication risk prediction system further comprises:

[0040] The variable importance ranking module is used to rank the importance of variables in the random forest model using the average prediction accuracy reduction value and the average Gini index reduction value respectively;

[0041] The model prediction effectiveness evaluation module is used to evaluate the overall prediction performance of the model through the model's prediction accuracy, sensitivity, and specificity; and to evaluate the model's prediction effect from three aspects: the model's discrimination, calibration, and clinical practicality.

[0042] As an implementation of the present invention, the prediction model building module performs hyperparameter tuning, and the parameter range is assumed to be: number of decision trees (ntree) = c (100, 200, 300, 400, 500); number of node partition features (mtry) = c (2, 3, 4, 5) during random forest modeling. The optimal model parameter configuration is finally obtained as follows: number of decision trees (ntree) = 200, number of node partition features (mtry) = 4, and the trained model has the best prediction performance.

[0043] According to another aspect of the present invention, the following technical solution is adopted: an electronic device includes 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 method when executing the computer program.

[0044] According to another aspect of the present invention, the following technical solution is adopted: a storage medium stores computer program instructions, and the computer program instructions implement the steps of the above method when executed by a processor.

[0045] The beneficial effects of the present invention are as follows: the method, system, electronic device and storage medium for constructing a vaginal delivery complication risk prediction model proposed in the present invention can more comprehensively predict the occurrence of various common complications of vaginal delivery, and objectively, quickly and effectively make a risk prediction assessment of the occurrence of complications based on the characteristic data of different pregnant women.

[0046] By applying the risk prediction model for common complications of vaginal delivery of the present invention, it is possible to quantify and estimate the occurrence of various common complications of vaginal delivery in pregnant women, so as to guide the formulation of targeted intervention plans, actively improve the safety of delivery for pregnant women, and improve their quality of life after delivery. It is of great significance to early identify high-risk pregnant women and take timely preventive intervention.

[0047] The present invention targets three common vaginal delivery complications (obstetric lacerations, postpartum hemorrhage, and retained placenta and fetal membranes without bleeding), weighs the comprehensive risk factors for occurrence and performs data analysis, and can predict the occurrence of a variety of common vaginal delivery complications. By more comprehensively evaluating the risk factors for common vaginal delivery complications, a wider range of vaginal delivery populations with risk factors are covered, which helps to achieve early identification and intervention of common vaginal delivery complications to ensure the safety of delivery for pregnant women.

[0048] The present invention introduces a machine learning algorithm in the field of artificial intelligence and constructs a risk prediction model for common complications of vaginal delivery based on the random forest algorithm. This model is different from the traditional risk prediction model based on the Logistic regression algorithm, and can better present the predictive performance of the risk of complications of vaginal delivery, thereby achieving a more accurate prediction of the occurrence of common complications of vaginal delivery.

[0049] Based on the exploration and analysis of risk factors for common complications of vaginal delivery, the present invention clearly defines relevant risk factors and uses a random forest algorithm to construct a comprehensive risk prediction model for common complications of vaginal delivery, which helps to provide a scientific means for formulating a systematic and standardized prevention and treatment plan for common complications of vaginal delivery. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 The present invention is a flowchart of a method for constructing a vaginal delivery complication risk prediction model in one embodiment of the present invention.

[0051] Figure 2 Schematic diagram of the composition of a system for building a vaginal delivery complication risk prediction model in one embodiment of the present invention.

[0052] Figure 3 FIG. 4 is a schematic diagram of the composition of an electronic device in one embodiment of the present invention. DETAILED DESCRIPTION

[0053] The preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0054] In order to further understand the present invention, preferred embodiments of the present invention are described below in conjunction with examples. However, it should be understood that these descriptions are only for further illustrating the features and advantages of the present invention, rather than limiting the claims of the present invention.

[0055] The description in this section is only for several typical embodiments, and the present invention is not limited to the scope of the embodiments. The same or similar prior art means and some technical features in the embodiments are mutually replaced within the scope of the present invention.

[0056] The description of the steps in each embodiment in the specification is only for the convenience of explanation, and the implementation method of the present application is not limited by the order of implementation of the steps.

[0057] The term “connection” in the specification includes both direct connection and indirect connection.

[0058] The present invention discloses a method for constructing a risk prediction model for vaginal delivery complications. Figure 1 This is a flow chart of a method for constructing a risk prediction model for vaginal delivery complications in one embodiment of the present invention; please refer to Figure 1 , the method for constructing a vaginal delivery complication risk prediction model comprises:

[0059] [Step S1] Obtaining maternal characteristic data;

[0060] [Step S2] Obtain risk indicators of maternal characteristic data;

[0061] [Step S3] Divide the maternal and child characteristic data into a training data set and a test data set for modeling;

[0062] [Step S4] Using the data set obtained in step S3, the divided data set is modeled in the manner of a random forest model.

[0063] In one embodiment of the present invention, step S4 includes:

[0064] The sample selection step is to select sample data from the set database;

[0065] a data preprocessing step, preprocessing the sample data selected by the sample selection unit;

[0066] Feature selection step, selecting feature data; the selected feature data include antenatal data (general condition of pregnant women, previous history of obstetrics and gynecology, and current pregnancy), delivery data (information of this delivery process, newborn conditions, and maternal conditions during delivery), and postpartum data (occurrence of common complications of vaginal delivery);

[0067] Variable selection step: select variable data; use Logistic regression and Lasso regression respectively to obtain the variables that have the most significant impact on the prediction results, and finally determine the best combination of predictive variables;

[0068] Modeling steps: establishing mathematical models; conducting mathematical modeling based on the best combination of predictive variables;

[0069] In the hyperparameter tuning step, a grid parameter search method is used to iterate through all possible parameter combinations to perform parameter combination tuning.

[0070] The variable importance ranking step uses the average prediction accuracy reduction value and the average Gini index reduction value to rank the importance of the variables in the random forest model;

[0071] The model prediction efficiency evaluation step evaluates the overall prediction performance of the model through the model's prediction accuracy, sensitivity, and specificity; and evaluates the model's prediction effect from three aspects: the model's discrimination, calibration, and clinical practicality.

[0072] In one use scenario of the present invention, in step S4, hyperparameter tuning is performed, and the parameter range for random forest modeling is assumed to be: number of decision trees (ntree) = c(100, 200, 300, 400, 500); number of node partition features (mtry) = c(2, 3, 4, 5). The optimal model parameter configuration is finally obtained as follows: number of decision trees (ntree) = 200, number of node partition features (mtry) = 4, and the trained model has the best prediction performance at this time. Of course, the above-mentioned variable importance ranking step and model prediction efficiency evaluation step can also be used as steps S5 and S6 in addition to step S4 of the method of the present invention.

[0073] In one embodiment of the present invention, the method for constructing a vaginal delivery complication risk prediction model may further include:

[0074] Step S5, performing variable importance analysis on the data set obtained in step S3;

[0075] The importance of variables in the random forest model was ranked using the average prediction accuracy reduction value and the average Gini index reduction value;

[0076] Step S6, evaluating the prediction performance of the model;

[0077] The overall prediction performance of the model was evaluated through its prediction accuracy, sensitivity and specificity. The prediction effect of the model was evaluated from three aspects: the discrimination, calibration and clinical practicality of the model.

[0078] The present invention further discloses a system for constructing a vaginal delivery complication risk prediction model. Figure 2 This is a schematic diagram of the components of a system for building a risk prediction model for vaginal delivery complications in one embodiment of the present invention; please refer to Figure 2 The vaginal delivery complication risk prediction model construction system includes: a feature data acquisition module 1, a risk indicator acquisition module 2, a data set allocation module 3 and a prediction model construction module 4.

[0079] The characteristic data acquisition module 1 is used to acquire the characteristic data of pregnant women; the risk indicator acquisition module 2 is used to acquire the risk indicators of the characteristic data of pregnant women; the data set allocation module 3 is used to divide the set characteristic data of pregnant women into a training data set and a test data set for modeling; the prediction model construction module 4 is used to model the divided data sets in the form of a random forest model to construct a vaginal delivery complication risk prediction model.

[0080] In one embodiment of the present invention, the prediction model building module 4 includes: a sample selection unit 41, a data preprocessing unit 42, a feature selection unit 43, a variable selection unit 44, a modeling unit 45 and a hyperparameter tuning unit 46.

[0081] The sample selection unit 41 is used to select sample data from a set database; the data preprocessing unit 42 is used to preprocess the sample data selected by the sample selection unit. The feature selection unit 43 is used to select feature data; the selected feature data include antenatal data (general condition of pregnant women, previous history of obstetrics and gynecology, and current pregnancy), intrapartum data (information of this labor process, newborn conditions, and maternal conditions during delivery) and postpartum data (common complications of vaginal delivery). The variable selection unit 44 is used to select variable data; Logistic regression and Lasso regression are used respectively to obtain the variables that have the most significant impact on the prediction results, and finally determine the best combination of predictive variables; the modeling unit 45 is used to establish a mathematical model and perform mathematical modeling based on the best combination of predictive variables. The hyperparameter tuning unit 46 is used to use a grid parameter search method to iterate through all possible parameter combinations to perform parameter combination tuning.

[0082] In one embodiment of the present invention, the vaginal delivery complication risk prediction system further includes: a variable importance ranking module 5 and a model prediction effectiveness evaluation module 6. The variable importance ranking module 5 is used to rank the importance of variables in the random forest model by using the predicted average precision reduction value and the average Gini index reduction value. The model prediction effectiveness evaluation module 6 is used to evaluate the overall prediction performance of the model through the prediction accuracy, sensitivity and specificity of the model; and to evaluate the model prediction effect from three aspects: the model's discrimination, calibration and clinical practicality.

[0083] In one use scenario of the present invention, the prediction model building module 4 performs hyperparameter tuning, and the parameter range is assumed to be: number of decision trees (ntree) = c (100, 200, 300, 400, 500); number of node partition features (mtry) = c (2, 3, 4, 5) during random forest modeling. The optimal model parameter configuration is finally obtained as follows: number of decision trees (ntree) = 200, number of node partition features (mtry) = 4, and the trained model has the best prediction performance.

[0084] The present invention also discloses an electronic device, Figure 3 FIG. 1 is a schematic diagram of the composition of an electronic device in an embodiment of the present invention; please refer to FIG. Figure 3 At the hardware level, the electronic device includes a memory, a processor and at least one network interface; the processor may be a microprocessor, and the memory may include a memory, such as a random access memory (RAM), or a non-volatile memory (non-volatile memory), etc. Of course, the electronic device may also be provided with other hardware as required.

[0085] The processor, network interface and memory may be interconnected via an internal bus, which may be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus may include an address bus, a data bus, a control bus, etc. The memory is used to store programs (which may include operating system programs and application programs); the program may include program code, which may include computer operation instructions. The memory may include internal memory and non-volatile memory, and provides instructions and data to the processor.

[0086] In one embodiment, the processor can read the corresponding program from the non-volatile memory into the memory and then run it; the processor can execute the program stored in the memory and is specifically used to perform the following operations (such as Figure 1 shown):

[0087] [Step S1] Obtaining maternal characteristic data;

[0088] [Step S2] Obtain risk indicators of maternal characteristic data;

[0089] [Step S3] Divide the set maternal and infant characteristic data into a training data set and a test data set for modeling;

[0090] [Step S4] Using the data set obtained in step S3, the divided data set is modeled in the manner of a random forest model.

[0091] [Step S5] Perform variable importance analysis on the data set obtained in step S3;

[0092] In one embodiment, the importance of variables in the random forest model is ranked using the average prediction accuracy reduction value and the average Gini index reduction value respectively;

[0093] [Step S6] Evaluate the prediction performance of the model;

[0094] In one embodiment, the overall prediction performance of the model is evaluated by the prediction accuracy, sensitivity, and specificity of the model; the prediction effect of the model is evaluated from three aspects: the discrimination, calibration, and clinical practicality of the model.

[0095] The present invention further discloses a storage medium having computer program instructions stored thereon, which, when executed by a processor, implements the following steps of the method of the present invention (eg Figure 1 shown):

[0096] [Step S1] Obtaining maternal characteristic data;

[0097] [Step S2] Obtain risk indicators of maternal characteristic data;

[0098] [Step S3] Divide the set maternal and infant characteristic data into a training data set and a test data set for modeling;

[0099] [Step S4] Using the data set obtained in step S3, the divided data set is modeled in the manner of a random forest model.

[0100] [Step S5] Perform variable importance analysis on the data set obtained in step S3;

[0101] In one embodiment, the importance of variables in the random forest model is ranked using the average prediction accuracy reduction value and the average Gini index reduction value respectively;

[0102] [Step S6] Evaluate the prediction performance of the model;

[0103] In one embodiment, the overall prediction performance of the model is evaluated by the prediction accuracy, sensitivity, and specificity of the model; the prediction effect of the model is evaluated from three aspects: the discrimination, calibration, and clinical practicality of the model.

[0104] By weighing comprehensive risk factors, the present invention analyzes the risk factors for three common vaginal delivery complications (obstetric lacerations, postpartum hemorrhage, and retained placenta and fetal membranes without bleeding), and on this basis, constructs a comprehensive prediction model for the risk of common complications of vaginal delivery, which is different from the risk prediction model for only a single complication. The new prediction model incorporates more potential risk factors and covers a wider range of people with vaginal delivery who have risk factors. It can more comprehensively evaluate the risk factors for common complications of vaginal delivery, which helps to achieve early identification and intervention of common complications of vaginal delivery to ensure the safety of delivery for pregnant women.

[0105] The present invention introduces a machine learning algorithm in the field of artificial intelligence and constructs a risk prediction model for common complications of vaginal delivery based on the random forest algorithm. This model is different from the traditional risk prediction model based on the Logistic regression algorithm, and can better present the prediction performance of the risk of complications of vaginal delivery.

[0106] Previous studies on complication risk prediction models were basically retrospective studies. This invention is the first to use a prospective study design to construct a prediction model for the risk of common complications of vaginal delivery, avoiding interference with research results due to partial confounding bias and better controlling data quality.

[0107] In summary, the method, system, electronic device and storage medium for constructing a vaginal delivery complication risk prediction model proposed in the present invention can more comprehensively predict the occurrence of various common complications of vaginal delivery, and objectively, quickly and effectively make a risk prediction assessment of the occurrence of complications based on the characteristic data of different pregnant women.

[0108] By applying the risk prediction model for common complications of vaginal delivery of the present invention, it is possible to quantify and estimate the occurrence of various common complications of vaginal delivery in pregnant women, so as to guide the formulation of targeted intervention plans, actively improve the safety of delivery for pregnant women, and improve their quality of life after delivery. It is of great significance to early identify high-risk pregnant women and take timely preventive intervention.

[0109] The present invention targets three common vaginal delivery complications (obstetric lacerations, postpartum hemorrhage, and retained placenta and fetal membranes without bleeding), weighs the comprehensive risk factors for occurrence and performs data analysis, and can predict the occurrence of a variety of common vaginal delivery complications. By more comprehensively evaluating the risk factors for common vaginal delivery complications, a wider range of vaginal delivery populations with risk factors are covered, which helps to achieve early identification and intervention of common vaginal delivery complications to ensure the safety of delivery for pregnant women.

[0110] The present invention introduces a machine learning algorithm in the field of artificial intelligence and constructs a risk prediction model for common complications of vaginal delivery based on the random forest algorithm. This model is different from the traditional risk prediction model based on the Logistic regression algorithm, and can better present the predictive performance of the risk of complications of vaginal delivery, thereby achieving a more accurate prediction of the occurrence of common complications of vaginal delivery.

[0111] It should be noted that the present application can be implemented in software and / or a combination of software and hardware; for example, it can be implemented using an application-specific integrated circuit (ASIC), a general-purpose computer, or any other similar hardware device. In some embodiments, the software program of the present application can be executed by a processor to implement the above steps or functions. Similarly, the software program of the present application (including related data structures) can be stored in a computer-readable recording medium; for example, a RAM memory, a magnetic or optical drive or a floppy disk and the like. In addition, some steps or functions of the present application can be implemented in hardware; for example, as a circuit that cooperates with a processor to perform various steps or functions.

[0112] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0113] The description and application of the present invention here are illustrative, and it is not intended to limit the scope of the present invention to the above-mentioned embodiments. The effects or advantages involved in the embodiments may not be embodied in the embodiments due to interference from various factors, and the description of the effects or advantages is not used to limit the embodiments. The deformation and change of the embodiments disclosed here are possible, and the replacement of the embodiments and the various equivalent parts are well known to those of ordinary skill in the art. It should be clear to those skilled in the art that the present invention can be implemented in other forms, structures, arrangements, proportions, and with other components, materials and parts without departing from the spirit or essential features of the present invention. Other deformations and changes can be made to the embodiments disclosed here without departing from the scope and spirit of the present invention.

Claims

1. A method for constructing a risk prediction model for vaginal delivery complications, characterized in that: The method for constructing a vaginal delivery complication risk prediction model comprises: Step S1, obtaining maternal characteristic data; Step S2, obtaining risk indicators of maternal and infant characteristic data; Step S3, dividing the set maternal and infant characteristic data into a training data set and a test data set for modeling; Step S4: Using the data set obtained in step S3, the divided data set is modeled in the manner of a random forest model.

2. The method for constructing a risk prediction model for vaginal delivery complications according to claim 1, characterized in that: The step S4 comprises: The sample selection step is to select sample data from the set database; a data preprocessing step, preprocessing the sample data selected by the sample selection unit; A feature selection step, selecting feature data; the selected feature data includes prenatal data, intrapartum data and postpartum data; Variable selection step: select variable data; use Logistic regression and Lasso regression respectively to obtain the variables that have the most significant impact on the prediction results, and finally determine the best combination of predictive variables; Modeling steps: establishing mathematical models; conducting mathematical modeling based on the best combination of predictive variables; In the hyperparameter tuning step, a grid parameter search method is used to iterate through all possible parameter combinations to perform parameter combination tuning. The variable importance ranking step uses the average prediction accuracy reduction value and the average Gini index reduction value to rank the importance of the variables in the random forest model; The model prediction efficiency evaluation step evaluates the overall prediction performance of the model through the model's prediction accuracy, sensitivity, and specificity; and evaluates the model's prediction effect from three aspects: the model's discrimination, calibration, and clinical practicality.

3. The method for constructing a risk prediction model for vaginal delivery complications according to claim 1, characterized in that: In step S4, hyperparameter tuning is performed, and the parameter range for random forest modeling is assumed to be: number of decision trees ntree = c(100, 200, 300, 400, 500); number of node partition features mtry = c(2, 3, 4, 5); and finally the best model parameter configuration is obtained as follows: number of decision trees ntree = 200, number of node partition features mtry = 4, at which time the trained model has the best prediction performance.

4. The method for constructing a risk prediction model for vaginal delivery complications according to claim 1, characterized in that: The method further comprises: Step S5, performing variable importance analysis on the data set obtained in step S3; The importance of variables in the random forest model was ranked using the average prediction accuracy reduction value and the average Gini index reduction value; Step S6, evaluating the prediction performance of the model; The overall prediction performance of the model was evaluated through its prediction accuracy, sensitivity and specificity. The prediction effect of the model was evaluated from three aspects: the discrimination, calibration and clinical practicality of the model.

5. A system for constructing a vaginal delivery complication risk prediction model, characterized in that: The vaginal delivery complication risk prediction model construction system comprises: A characteristic data acquisition module is used to obtain characteristic data of pregnant women; A risk indicator acquisition module is used to obtain risk indicators of maternal and infant characteristic data; A data set allocation module is used to divide the set maternal and infant characteristic data into a training data set and a test data set for modeling; The prediction model building module is used to model the divided data set in the form of a random forest model to build a vaginal delivery complication risk prediction model.

6. The system for constructing a vaginal delivery complication risk prediction model according to claim 5, characterized in that: The prediction model building module includes: A sample selection unit, used to select sample data from a set database; A data preprocessing unit, used for preprocessing the sample data selected by the sample selection unit; A feature selection unit, used to select feature data; the selected feature data includes prenatal data, delivery data and postnatal data; The variable selection unit is used to select variable data; Logistic regression and Lasso regression are used to obtain the variables that have the most significant impact on the prediction results, and finally determine the best combination of predictive variables; A modeling unit is used to establish a mathematical model and perform mathematical modeling according to the best combination of predictive variables; The hyperparameter tuning unit is used to use a grid parameter search method to iterate through all possible parameter combinations to perform parameter combination tuning.

7. The system for constructing a vaginal delivery complication risk prediction model according to claim 6, characterized in that: The vaginal delivery complication risk prediction system further comprises: The variable importance ranking module is used to rank the importance of variables in the random forest model using the average prediction accuracy reduction value and the average Gini index reduction value respectively; The model prediction effectiveness evaluation module is used to evaluate the overall prediction performance of the model through the model's prediction accuracy, sensitivity, and specificity; and to evaluate the model's prediction effect from three aspects: the model's discrimination, calibration, and clinical practicality.

8. The vaginal delivery complication risk prediction system according to claim 5, characterized in that: The prediction model construction module performs hyperparameter tuning, and the parameter range is assumed to be: number of decision trees ntree=c(100, 200, 300, 400, 500) when random forest modeling; number of node partition features mtry=c(2, 3, 4, 5); the optimal model parameter configuration is finally obtained as: number of decision trees ntree=200, number of node partition features mtry=4, and the trained model has the best prediction performance at this time.

9. An electronic 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 method according to any one of claims 1 to 4 are implemented.

10. A storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images