Gynaecology and obstetrics pregnant woman evaluation analysis method and system based on community antenatal care data flow

Abstract

The invention provides an obstetrics and gynecology pregnant woman evaluation analysis method and system based on community antenatal care data flow, and relates to the technical field of medical health analysis, the obstetrics and gynecology pregnant woman evaluation analysis method based on community antenatal care data flow comprises the following steps: constructing a community antenatal care data flow storage platform, and obtaining and storing antenatal care data in real time; based on the acquired antenatal care data, constructing a coupling matrix through antenatal care period decomposition and risk factor decomposition; based on the constructed coupling matrix, utilizing an antenatal care risk tree analysis method to establish an antenatal care risk tree risk model, and performing risk assessment on an antenatal care result through the antenatal care risk tree risk model; and displaying a risk assessment result through a data visualization technology. According to the invention, risk assessment is carried out on the antenatal care result through the antenatal care risk tree risk model, and a more accurate and personalized assessment result can be provided.

Description

Technical Field

[0001] The present invention relates to the field of medical health analysis technology, and in particular to an obstetrics and gynecology pregnant woman assessment and analysis method and system based on community prenatal examination data stream. Background Art

[0002] Prenatal check-ups are an important means of obtaining the health status of pregnant women and their fetuses. Reasonable prenatal check-ups can not only help pregnant women understand the health status of themselves and their fetuses, but also adjust their living habits, diet and exercise plans in a timely manner based on the prenatal check-up information. Usually, when pregnant women learn of their pregnancy, they will create a file at a community hospital and receive regular prenatal check-ups. These regular check-ups help identify and respond to potential health problems early. In a community medical environment, this routine prenatal check-up activity generates a large amount of data, including basic health indicators of pregnant women, fetal development, genetic disease screening results, etc., forming a multi-dimensional community prenatal check-up data stream. Through in-depth analysis and evaluation of these data, medical experts can assess pregnant women for high-risk pregnancies, predict possible health problems, and provide personalized medical advice and intervention measures accordingly.

[0003] At present, data fragmentation and information islands are common in traditional prenatal data management. Data between different medical institutions are difficult to integrate, which limits the comprehensiveness and accuracy of medical decisions. The inconsistency of data makes it impossible for medical providers to obtain complete health records of pregnant women, affecting the efficiency and effectiveness of medical services. In addition, when processing prenatal data, the comprehensive analysis of periodicity and risk factors is often ignored, which limits the comprehensive understanding of medical service providers of the dynamic changes in the health status of pregnant women. In addition, the lack of sophisticated data analysis tools makes it difficult to extract information that is instructive for clinical practice from large amounts of data.

[0004] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0005] In view of this, the present invention provides an obstetrics and gynecology pregnant woman assessment and analysis method and system based on community prenatal examination data stream to solve the above-mentioned problems.

[0006] In order to solve the above problems, the specific technical solutions adopted by the present invention are as follows: According to one aspect of the present invention, a method for evaluating and analyzing obstetrics and gynecology pregnant women based on community prenatal examination data stream is provided. The method for evaluating and analyzing obstetrics and gynecology pregnant women based on community prenatal examination data stream comprises the following steps: S1. Build a community prenatal examination data stream storage platform to obtain and store prenatal examination data in real time; S2. Based on the acquired prenatal examination data, a coupling matrix is ​​constructed by decomposing the prenatal examination cycle and risk factors; S3. Based on the constructed coupling matrix, a prenatal examination risk tree risk model is established using the prenatal examination risk tree analysis method, and the prenatal examination risk tree risk model is used to conduct risk assessment on the prenatal examination results; S4. Display risk assessment results through data visualization technology; Preferably, based on the acquired prenatal examination data, constructing a coupling matrix through prenatal examination cycle decomposition and risk factor decomposition includes the following steps: S21, preprocessing the acquired prenatal examination data, including data cleaning and standardization; S22, constructing a prenatal examination data classification model, and classifying the prenatal examination data after prenatal examination through the prenatal examination data classification model; S23, performing time series analysis on each type of data in the prenatal examination data, and according to the analysis results, performing period decomposition on the prenatal examination data using a preset period standard; S24. Construct a cycle decomposition structure according to the cycle decomposition results and classification results of the prenatal examination data; S25. According to the inherent attributes of each type of data in the prenatal examination data, identify the risk factors of each type of data, and construct a risk factor decomposition structure through step-by-step decomposition and analysis; S26. Establish a coupling matrix based on the cycle decomposition structure and risk factor decomposition structure.

[0007] Preferably, constructing a prenatal examination data classification model, and classifying the prenatal examination data after prenatal examination by the prenatal examination data classification model comprises the following steps: S221, collect historical prenatal examination sample data, and establish a training data set through feature selection processing; S222. Construct and train a prenatal examination data classification model using a support vector machine algorithm based on the obtained training data set; S223. The preprocessed antenatal examination data is used as input and input into the trained antenatal examination data classification model to classify the preprocessed antenatal examination data.

[0008] Preferably, collecting historical prenatal examination sample data and establishing a training data set through feature selection processing includes the following steps: S2211. Perform dimensionality reduction processing on historical prenatal examination sample data by using dimensionality reduction technology; S2212, performing correlation feature selection on the historical prenatal examination sample data after dimension reduction to establish a feature data set; S2213. Optimize the feature data set using a spectral clustering algorithm to obtain a new feature data set, and use the new feature data set as a training data set.

[0009] Preferably, processing the feature data set using a spectral clustering algorithm to obtain a new feature data set includes the following steps: S22131. For each feature data point in the feature data set, calculate the similarity between each pair of feature data points; S22132, constructing an undirected graph using each feature data point in the feature data set as a node, and using the similarity between the feature data points as the weight of the edge between the nodes in the undirected graph; S22133, constructing a weighted connection matrix and a degree matrix according to the similarity between nodes and the sum of the weights of all edges connected to each node; S22134. Calculate a normalized Laplace matrix based on the constructed weighted connection matrix and degree matrix; S22135, performing eigenvalue decomposition on the calculated normalized Laplace matrix, and extracting eigenvectors corresponding to several minimum non-zero eigenvalues ​​therein; S22136. Divide the data points into several clusters according to the feature vectors, and select the cluster center from each cluster as a representative point to form a new feature data set.

[0010] Preferably, the calculation formula for calculating the normalized Laplace matrix based on the constructed weighted connection matrix and degree matrix is: R=F -1 / 2 W F -1 / 2 ; Where R represents the normalized Laplace matrix; F represents the degree matrix; F -1 / 2 represents the diagonal matrix formed by taking the square root of the reciprocal of each diagonal element of the degree matrix F; W represents the weighted connection matrix.

[0011] Preferably, performing time series analysis on each type of data in the prenatal examination data, and performing period decomposition on the prenatal examination data according to a preset period standard according to the analysis result comprises the following steps: S231, obtaining the real-time acquisition time of the prenatal examination data, and arranging each type of data in the prenatal examination data in the order of the real-time acquisition time to form a time series; S232, dividing each type of data in the prenatal examination data into periods according to a preset period standard; S233, grouping and arranging the prenatal examination data after the cycle division according to the cycle.

[0012] Preferably, based on the constructed coupling matrix, a prenatal examination risk tree risk model is established using a prenatal examination risk tree analysis method, and risk assessment of prenatal examination results using the prenatal examination risk tree risk model includes the following steps: S31, determine the top event of the prenatal examination risk tree, use the coupling matrix to analyze the top event of the prenatal examination risk tree, identify the basic events and build a prenatal examination risk tree risk model; S32. Calculate the probability of occurrence of the top event by probability calculation method according to the antenatal examination risk tree risk model; S33. Calculate the sensitivity coefficient of each basic event through sensitivity analysis according to the probability of occurrence of the top event; S34. Conduct risk assessment on prenatal examination results based on sensitivity analysis results.

[0013] Preferably, according to the prenatal examination risk tree risk model, calculating the probability of occurrence of the top event by a probability calculation method includes the following steps: S321, assigning a basic probability value to each basic event in the antenatal examination risk tree risk model; S322, determining the logical relationship between events according to the logic gates in the antenatal examination risk tree risk model; S323. Starting from the basic event at the bottom of the antenatal examination risk tree risk model, calculate upward step by step according to the logical relationship to obtain the probability of the top event occurring.

[0014] Preferably, according to the probability of occurrence of the top event, the calculation formula for calculating the sensitivity coefficient of each basic event through sensitivity analysis is: ; In the formula, H i Represents basic events i The sensitivity coefficient of Indicates top events T The partial derivative value of the probability value of occurrence; D ( T ) indicates the top event T Probability value of occurrence; Represents basic events i The partial derivative value of the basic probability value of ; k i Represents basic events i The basic probability value of .

[0015] According to another aspect of the present invention, a system for evaluating and analyzing obstetrics and gynecology pregnant women based on community prenatal examination data stream is provided, and the system for evaluating and analyzing obstetrics and gynecology pregnant women based on community prenatal examination data stream comprises: a data existence module, a matrix construction module, a risk assessment module and a result display module, wherein the data existence module, the matrix construction module, the risk assessment module and the result display module are connected in sequence; The data storage module is used to build a community prenatal examination data stream storage platform to obtain and store prenatal examination data in real time; A matrix construction module is used to construct a coupling matrix based on the acquired prenatal examination data through prenatal examination cycle decomposition and risk factor decomposition; The risk assessment module is used to establish a prenatal examination risk tree risk model based on the constructed coupling matrix using the prenatal examination risk tree analysis method, and to conduct risk assessment on the prenatal examination results through the prenatal examination risk tree risk model; The result display module is used to display the risk assessment results through data visualization technology.

[0016] The beneficial effects of the present invention are: 1. The present invention constructs a community prenatal examination data stream storage platform, which can acquire and store prenatal examination data in real time to ensure the timeliness and accuracy of the data. Based on the acquired prenatal examination data, a coupling matrix is ​​constructed by decomposing the prenatal examination cycle and the risk factors, which helps to systematically identify and analyze the key information and risk factors in the prenatal examination data. The coupling matrix can clearly display the relationship and interaction between different prenatal examination cycles and risk factors, and provide strong support for subsequent risk assessment. The prenatal examination risk tree analysis method is used to establish a prenatal examination risk tree risk model, which can intuitively display various risks and abnormal situations that may exist in the prenatal examination process. The prenatal examination risk tree risk model is used to conduct risk assessment on the prenatal examination results, which can provide more accurate and personalized assessment results.

[0017] 2. The present invention collects historical prenatal examination sample data and uses feature selection processing and dimensionality reduction technology to construct an efficient and accurate prenatal examination data classification model. The spectral clustering algorithm is used to optimize the feature data set to obtain a more compact and representative feature set, thereby improving the efficiency and accuracy of the classification model. The calculation of the standardized Laplace matrix and the eigenvalue decomposition can ensure the reasonable clustering of data points, thereby extracting the most distinctive features. The constructed coupling matrix can clearly display the period decomposition structure and risk factor decomposition structure of the prenatal examination data, which is convenient for rapid identification and understanding of key information.

[0018] 3. The present invention can systematically evaluate the risks that pregnant women may face throughout the entire prenatal examination process through the prenatal examination risk tree analysis method, which helps doctors to fully understand the health status and potential problems of pregnant women, so as to formulate more accurate and effective treatment and management plans. The prenatal examination risk tree risk model intuitively displays various risk factors and their relationships with each other in a tree structure, and can clearly see which factors may have the greatest impact on the health of pregnant women and fetuses. The probability of occurrence of top events (i.e., adverse pregnancy outcomes) is calculated through probability calculation methods, providing doctors with quantitative risk assessment results, and they can judge the severity of the risk based on specific values, so as to make more scientific decisions, which is helpful to improve the health level of mothers and babies. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings: Figure 1 is a flow chart of a method for evaluating and analyzing obstetrics and gynecology pregnant women based on community prenatal examination data stream according to an embodiment of the present invention; Figure 2 It is a principle block diagram of an obstetrics and gynecology pregnant woman assessment and analysis system based on community prenatal examination data stream according to an embodiment of the present invention.

[0020] In the figure: 1. Data existence module; 2. Matrix construction module; 3. Risk assessment module; 4. Result display module. DETAILED DESCRIPTION

[0021] In order to enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this application.

[0022] According to an embodiment of the present invention, a method and system for evaluating and analyzing obstetrics and gynecology pregnant women based on community prenatal examination data streams are provided.

[0023] The present invention will now be further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, according to one embodiment of the present invention, a method for evaluating and analyzing obstetrics and gynecology pregnant women based on community prenatal examination data stream is provided. The method for evaluating and analyzing obstetrics and gynecology pregnant women based on community prenatal examination data stream comprises the following steps: S1. Build a community prenatal examination data stream storage platform to obtain and store prenatal examination data in real time; Specifically, the community prenatal examination data stream storage platform should be divided into the front-end display layer, the back-end service layer and the data storage layer. The front-end display layer is responsible for the display and interaction of the user interface, the back-end service layer handles business logic and data processing, and the data storage layer is responsible for storing user data and platform information. Prenatal examination data can be collected through various channels such as hospital systems, medical equipment, and personal input by pregnant women.

[0024] Prenatal examination data includes basic personal information, medical history, physical examination data and test results at various stages of pregnancy, etc.

[0025] S2. Based on the acquired prenatal examination data, a coupling matrix is ​​constructed by decomposing the prenatal examination cycle and risk factors; As a preferred implementation, based on the acquired prenatal examination data, constructing a coupling matrix through prenatal examination cycle decomposition and risk factor decomposition includes the following steps: S21, preprocessing the acquired prenatal examination data, including data cleaning and standardization; It should be noted that data cleaning mainly solves errors and inconsistencies in data, including removing duplicate records, handling missing values, and correcting errors and outliers. Standardization is to adjust data to a unified format and range for easy analysis and processing.

[0026] S22, constructing a prenatal examination data classification model, and classifying the prenatal examination data after prenatal examination through the prenatal examination data classification model; As a preferred implementation, constructing a prenatal examination data classification model, and classifying the prenatal examination data after prenatal examination by the prenatal examination data classification model includes the following steps: S221, collect historical prenatal examination sample data, and establish a training data set through feature selection processing; As a preferred implementation, collecting historical prenatal examination sample data and establishing a training data set through feature selection processing includes the following steps: S2211. Perform dimensionality reduction processing on historical prenatal examination sample data by using dimensionality reduction technology; Specifically, dimensionality reduction techniques include principal component analysis (PCA), linear discriminant analysis, t-distributed random neighbor embedding, and so on.

[0027] For example, using principal component analysis (PCA) as an example: Normalize data: Before applying PCA, it is usually necessary to normalize the data to eliminate the dimensional differences between different features.

[0028] Calculate Covariance Matrix: Calculate the covariance matrix or correlation coefficient matrix of the standardized data.

[0029] Calculate the eigenvalues ​​and eigenvectors of the covariance matrix: These eigenvalues ​​and eigenvectors describe the main directions of variation of the data.

[0030] Select principal components: Select the first n principal components based on the size of the eigenvalues, which will be used as the new feature space. Usually, you can select the first n principal components whose cumulative eigenvalues ​​account for a certain threshold (such as 90%).

[0031] Transform data: Transform the data from the original feature space to the new principal component space, usually by performing a matrix multiplication of the original data with the eigenvectors of the selected principal components.

[0032] S2212, performing correlation feature selection on the historical prenatal examination sample data after dimension reduction to establish a feature data set; It should be noted that the correlation feature selection of the historical prenatal examination sample data after dimensionality reduction is aimed at identifying and retaining the features most relevant to the target variable. The Pearson correlation coefficient method is used to calculate the correlation between the features after dimensionality reduction and the target variable. According to the results of the correlation analysis, a correlation threshold is set. For example, features with a correlation with the target variable higher than a certain threshold can be selected. The selected features and the target variable are combined into a feature data set.

[0033] S2213. Optimize the feature data set using a spectral clustering algorithm to obtain a new feature data set, and use the new feature data set as a training data set.

[0034] As a preferred implementation, the feature data set is processed using a spectral clustering algorithm to obtain a new feature data set, including the following steps: S22131. For each feature data point in the feature data set, calculate the similarity between each pair of feature data points; It should be noted that when calculating the similarity between each pair of feature data points in the feature data set, the Euclidean distance is used for similarity measurement.

[0035] S22132, constructing an undirected graph using each feature data point in the feature data set as a node, and using the similarity between the feature data points as the weight of the edge between the nodes in the undirected graph; Specifically, each feature data point in the feature data set is used as a node to construct an undirected graph, and the similarity between the feature data points is used as the weight of the edge between the nodes in the undirected graph, including the following steps: Create an empty undirected graph whose node set is all the data points in the feature dataset; The similarity between each pair of data points in the feature dataset is stored in a similarity matrix; Traverse the similarity matrix, and for each non-zero element in the matrix, add an edge from node to node in the undirected graph, setting the weight of this edge to the similarity value.

[0036] S22133, constructing a weighted connection matrix and a degree matrix according to the similarity between nodes and the sum of the weights of all edges connected to each node; It should be noted that the degree matrix is ​​a diagonal matrix where each diagonal element represents the degree of a data point, that is, the sum of the similarities between the data point and all other data points.

[0037] S22134. Calculate a normalized Laplace matrix based on the constructed weighted connection matrix and degree matrix; Specifically, the calculation formula for calculating the standardized Laplace matrix based on the constructed weighted connection matrix and degree matrix is: R=F -1 / 2 W F -1 / 2 ; Where R represents the normalized Laplace matrix; F represents the degree matrix; F -1 / 2 represents the diagonal matrix formed by taking the square root of the reciprocal of each diagonal element of the degree matrix F; W represents the weighted connection matrix.

[0038] S22135, performing eigenvalue decomposition on the calculated normalized Laplace matrix, and extracting eigenvectors corresponding to several minimum non-zero eigenvalues ​​therein; Specifically, eigenvalue decomposition is a mathematical process that decomposes a matrix into its eigenvalues ​​and eigenvectors. When performing eigenvalue decomposition, a series of eigenvalues ​​and corresponding eigenvectors are obtained. Since the normalized Laplace matrix is ​​semi-positive definite, its eigenvalues ​​are all non-negative. The specific steps to extract the eigenvectors corresponding to several minimum non-zero eigenvalues ​​are as follows: Perform eigenvalue decomposition on the standardized Laplace matrix to obtain a set of eigenvalues ​​and corresponding eigenvectors; Sort the eigenvalues ​​from smallest to largest and adjust the order of the eigenvectors accordingly; Skip zero eigenvalues ​​(if any) as they do not contain any information about the clustering of the data; Select several smallest non-zero eigenvalues ​​(usually a number close to the expected number of clusters) and extract their corresponding eigenvectors.

[0039] For example, suppose there is a data set containing 100 data points, and a normalized Laplacian matrix has been constructed based on the similarity between these data points, and it is hoped that these data points will be clustered into 5 categories; Decompose the normalized Laplace matrix L using the eigenvalue decomposition algorithm; two arrays will be returned: one is an array of eigenvalues ​​in ascending order, and the other is an array of eigenvectors corresponding to these eigenvalues. It is assumed that the eigenvalue array obtained after eigenvalue decomposition is: [0, 0.05, 0.1, 0.15, 0.2, 0.25, 0.3, ...]; Skip zero eigenvalues ​​(if any) because they do not provide useful information about data clustering. Extract the eigenvectors corresponding to the smallest 5 non-zero eigenvalues ​​(0.05 to 0.25 in this example) from the eigenvalue array. These eigenvectors will form a new matrix V, where each row represents the representation of a data point in the new feature space. Cluster the rows of the matrix V. Since 5 eigenvectors were selected, the clustering algorithm will try to divide the data points into 5 categories.

[0040] S22136. Divide the data points into several clusters according to the feature vectors, and select the cluster center from each cluster as a representative point to form a new feature data set.

[0041] It should be noted that the selected feature vector is used as a new feature to map the data points into this low-dimensional feature space. The coordinates of each data point in the new feature space are composed of its corresponding feature vector value. In the new feature space, a clustering algorithm such as K-means is used to divide the data points into k clusters. For each cluster, its cluster center is selected as the representative point. The cluster center is usually the mean of all data points in the cluster.

[0042] S222. Construct and train a prenatal examination data classification model using a support vector machine algorithm based on the obtained training data set; Specifically, based on the obtained training data set, constructing and training the prenatal examination data classification model through the support vector machine algorithm includes the following steps: Use the support vector machine algorithm to build a classification model and select an appropriate kernel function (such as linear, polynomial, radial basis function, etc.); Adjust the model's hyperparameters, such as the penalty coefficient C (which controls the degree of penalty for misclassification) and the coefficients of the kernel function (such as the gamma parameter of the RBF kernel); The classification model is trained using the training data set, and the performance of the model is evaluated by obtaining the test data set to obtain the prenatal examination data classification model.

[0043] S223. The preprocessed antenatal examination data is used as input and input into the trained antenatal examination data classification model to classify the preprocessed antenatal examination data.

[0044] S23, performing time series analysis on each type of data in the prenatal examination data, and according to the analysis results, performing period decomposition on the prenatal examination data using a preset period standard; As a preferred implementation, performing time series analysis on each type of data in the prenatal examination data, and decomposing the prenatal examination data into periods according to the analysis results by using a preset period standard includes the following steps: S231, obtaining the real-time acquisition time of the prenatal examination data, and arranging each type of data in the prenatal examination data in the order of the real-time acquisition time to form a time series; It should be noted that the real-time acquisition time (usually date or timestamp) of each piece of data is extracted from the prenatal examination data set, and each piece of data in the prenatal examination data (such as blood pressure, blood sugar, weight, etc.) is sorted in chronological order according to these timestamps. The sorted data forms a time series, which shows the changes of each prenatal examination data over time.

[0045] S232, dividing each type of data in the prenatal examination data into periods according to a preset period standard; It should be noted that each type of data in the prenatal examination data is divided into cycles according to the pregnancy cycle standard. The pregnancy cycle standard is to group the data according to the pregnancy stage of the pregnant woman (such as early pregnancy, mid-pregnancy, late pregnancy) or the specific number of weeks of pregnancy (such as every 4 weeks as a cycle). According to the number of weeks of pregnancy, the prenatal examination data can be divided into the corresponding cycle.

[0046] S233, grouping and arranging the prenatal examination data after the cycle division according to the cycle.

[0047] S24. Construct a cycle decomposition structure according to the cycle decomposition results and classification results of the prenatal examination data; S25. According to the inherent attributes of each type of data in the prenatal examination data, identify the risk factors of each type of data, and construct a risk factor decomposition structure through step-by-step decomposition and analysis; S26. Establish a coupling matrix based on the cycle decomposition structure and risk factor decomposition structure.

[0048] It should be noted that the decomposition of the periodicity of prenatal examination data and the decomposition of risk factors adopts the WBS-RBS (WorkBreakdown Structure - Risk Breakdown Structure) decomposition structure principle. Using the WBS-RBS decomposition structure principle to handle the periodicity of prenatal examination data and the decomposition of risk factors is a structured and systematic method. This method can more clearly understand the various stages and potential risk factors in the prenatal examination cycle.

[0049] WBS is a method for breaking down complex projects into smaller, more manageable parts. In the context of prenatal examination data, WBS can divide the entire pregnancy into different examination stages, each of which includes different examination items and objectives; RBS is a method for identifying and classifying project risks. It can classify risks into different categories and further refine them to specific risk sources. In prenatal examinations, potential health risks related to pregnancy can be identified and managed.

[0050] S3. Based on the constructed coupling matrix, a prenatal examination risk tree risk model is established using the prenatal examination risk tree analysis method, and the prenatal examination risk tree risk model is used to conduct risk assessment on the prenatal examination results; It should be noted that the main idea of ​​the prenatal risk tree analysis method is the fault tree analysis principle, which aims to depict the directed logic tree of the accident through logical deduction, so as to deeply analyze the phenomenon, cause and result of the accident, and then provide effective measures to prevent accidents. The prenatal risk tree analysis method is based on the fault tree principle. By constructing a logically clear fault tree, it deeply analyzes the risk factors and their interrelationships in the prenatal examination process, and provides effective measures and basis for preventing adverse pregnancy outcomes.

[0051] As a preferred implementation, based on the constructed coupling matrix, a prenatal examination risk tree risk model is established using a prenatal examination risk tree analysis method, and risk assessment of prenatal examination results is performed using the prenatal examination risk tree risk model, including the following steps: S31, determine the top event of the prenatal examination risk tree, use the coupling matrix to analyze the top event of the prenatal examination risk tree, identify the basic events and build a prenatal examination risk tree risk model; It should be noted that the top event is the most critical component in the antenatal risk tree analysis. It represents the focus of the analysis and is usually a major adverse outcome that is not expected to occur. In the context of antenatal examinations, a typical top event may be "serious complications during antenatal examinations."

[0052] The coupling matrix connects the period decomposition structure and the risk factor decomposition structure, providing the necessary input for the prenatal risk tree analysis. This step involves the following operations: Analyze the coupling matrix: Check the coupling degree of each risk factor related to the top event and determine the key factors affecting the top event.

[0053] Select basic events: Based on the analysis of the coupling matrix, select those risk factors that have the greatest impact on the top event as the basic events of the prenatal risk tree. For example, if "hypertension" and "diabetes" are highly correlated with the top event in the coupling matrix, these factors will be selected as basic events.

[0054] Among them, after determining the top event and related basic events, building a prenatal risk tree model involves the following steps: Determine the structure of the prenatal risk tree: Start from the top event and expand downward to include all identified basic events. Use "AND gate" and "OR gate" to represent the logical relationship between events: AND Gate: All input events must occur simultaneously for the output event to occur.

[0055] OR gate: The occurrence of any input event can lead to the occurrence of an output event.

[0056] Connect logic gates: Select appropriate logic gates to connect top events and basic events based on the dependencies and interactions between events. For example, if the top event "serious complications during prenatal checkup" is caused by "hypertension" and "diabetes", these two basic events are connected to the top event through an "AND gate".

[0057] S32. Calculate the probability of occurrence of the top event by probability calculation method according to the antenatal examination risk tree risk model; As a preferred implementation, according to the antenatal examination risk tree risk model, calculating the probability of occurrence of the top event by a probability calculation method includes the following steps: S321, assigning a basic probability value to each basic event in the antenatal examination risk tree risk model; It should be noted that the assignment of basic probability values ​​is usually based on historical data, expert opinions, or related research. For example, based on historical data, past medical records are reviewed and the frequency of related complications is statistically analyzed. The basic probability value assigns a probability value to each basic event, usually between 0 (not occurring) and 1 (definitely occurring).

[0058] The hypothetical antenatal risk tree includes the following basic events: Hypertension: Historical data indicate that the incidence among pregnant women is approximately 10%, so a probability value of 0.1 is assigned.

[0059] Diabetes: According to the study, the incidence of diabetes during pregnancy is about 7%, so the probability value is assigned 0.07.

[0060] Lifestyle risks (such as smoking and poor diet): Experts estimate the impact of this risk to be about 5%, assigning a probability value of 0.05.

[0061] Environmental risks (such as exposure to hazardous substances): more difficult to quantify accurately, with an estimated probability of 3% and an assigned probability value of 0.03.

[0062] S322, determining the logical relationship between events according to the logic gates in the antenatal examination risk tree risk model; It should be noted that in the prenatal risk tree analysis (FTA), the logical relationship between events is determined by using logic gates, mainly including "AND gates" and "OR gates". These logic gates define how different events interact with each other to cause the occurrence of the top event.

[0063] In the prenatal check-up risk tree, if the top event "serious complications during prenatal check-up" requires multiple conditions to be met at the same time (for example, hypertension and diabetes exist at the same time), these conditions are connected through an AND gate; if the top event can be caused by multiple different single conditions, such as "serious complications during prenatal check-up" can be caused independently by either "hypertension" or "lifestyle risks", these events are connected through an OR gate.

[0064] S323. Starting from the basic event at the bottom of the antenatal examination risk tree risk model, calculate upward step by step according to the logical relationship to obtain the probability of the top event occurring.

[0065] It should be noted that each basic event has a pre-assigned probability value that reflects the probability of its independent occurrence. For example, the probability of hypertension is 0.110%), the probability of diabetes is 0.07 (7%), the probability of lifestyle risk is 0.05 (5%), and the probability of environmental risk is 0.03 (3%); The AND gate is used to calculate the probability of two events occurring simultaneously. For example, the probability of hypertension and diabetes occurring simultaneously is 0.007 (0.7%), which is obtained by multiplying the probabilities of the two events together.

[0066] The OR gate is used to calculate the total probability of any event occurring. This includes the probability of the individual events plus the probability of any combination, minus the probability of them occurring together.

[0067] Combine the results of the AND gate and the OR gate to calculate the total probability of the top event "serious complications during prenatal check-up". The specific calculation is as follows: First add up the probabilities of all basic events.

[0068] Then subtract the overlap in probability between the events when they are combined pairwise.

[0069] Finally, if three or more events occur together, add the probability of that part.

[0070] The calculation results show that the probability of the top event is approximately 0.085 or 8.5, which means that under all considered risk factors, there is an 8.5% probability of serious complications.

[0071] S33. Calculate the sensitivity coefficient of each basic event through sensitivity analysis according to the probability of occurrence of the top event; As a preferred implementation, according to the probability of occurrence of the top event, the calculation formula for calculating the sensitivity coefficient of each basic event through sensitivity analysis is: ; In the formula, H i Represents basic events i The sensitivity coefficient of Indicates top events T The partial derivative value of the probability value of occurrence; D ( T ) indicates the top event T Probability value of occurrence; Represents basic events i The partial derivative value of the basic probability value of ; k i Represents basic events i The basic probability value of .

[0072] S34. Conduct risk assessment on prenatal examination results based on sensitivity analysis results.

[0073] Specifically, based on the results of sensitivity analysis, the risk assessment of prenatal examination results includes the following steps: Identify key risk factors: Basic events with high sensitivity coefficients mean that they have a greater impact on the probability of the top event occurring. These factors should be the focus of risk management.

[0074] Develop risk mitigation strategies: Based on the results of the sensitivity analysis, develop targeted risk mitigation measures, especially for those basic events with high sensitivity coefficients.

[0075] Optimal resource allocation: Allocate resources based on sensitivity coefficients, giving priority to basic events that have the greatest impact on the probability of top events.

[0076] S4. Display risk assessment results through data visualization technology; Specifically, the risk assessment results can be displayed using data visualization technology using bar charts, which can effectively display the sensitivity coefficient of each basic event. The length of each bar represents the sensitivity of the corresponding event, making it clear at a glance which factors have the greatest impact on the probability of the top event.

[0077] like Figure 2 As shown, according to another embodiment of the present invention, a gynecology and obstetrics pregnant woman assessment and analysis system based on community prenatal examination data stream is provided, and the gynecology and obstetrics pregnant woman assessment and analysis system based on community prenatal examination data stream includes: a data existence module 1, a matrix construction module 2, a risk assessment module 3 and a result display module 4, wherein the data existence module 1, the matrix construction module 2, the risk assessment module 3 and the result display module 4 are connected in sequence; Data storage module 1 is used to build a community prenatal examination data stream storage platform to obtain and store prenatal examination data in real time; The matrix construction module 2 is used to construct a coupling matrix based on the acquired prenatal examination data by decomposing the prenatal examination cycle and the risk factors; The risk assessment module 3 is used to establish a prenatal examination risk tree risk model based on the constructed coupling matrix by using the prenatal examination risk tree analysis method, and to conduct risk assessment on the prenatal examination results by using the prenatal examination risk tree risk model; The result display module 4 is used to display the risk assessment results through data visualization technology.

[0078] To sum up, with the help of the above-mentioned technical scheme of the present invention, the present invention can acquire and store prenatal examination data in real time by constructing a community prenatal examination data stream storage platform to ensure the timeliness and accuracy of the data. Based on the acquired prenatal examination data, a coupling matrix is ​​constructed by decomposing the prenatal examination cycle and the risk factors, which helps to systematically identify and analyze the key information and risk factors in the prenatal examination data. The coupling matrix can clearly display the relationship and interaction between different prenatal examination cycles and risk factors, and provide strong support for subsequent risk assessment. The prenatal examination risk tree risk model is established by using the prenatal examination risk tree analysis method, which can intuitively display various risks and abnormal situations that may exist in the prenatal examination process. The risk assessment of the prenatal examination results is performed through the prenatal examination risk tree risk model, which can provide more accurate and personalized assessment results. The present invention collects historical antenatal examination sample data, utilizes feature selection processing and dimensionality reduction technology, can construct an efficient and accurate antenatal examination data classification model, utilizes spectral clustering algorithm to optimize feature data set, can obtain more compact and representative feature set, improves efficiency and accuracy of classification model, and the calculation of standardized Laplace matrix and eigenvalue decomposition can ensure reasonable clustering of data points, thereby extracting the most distinguishing features, and the constructed coupling matrix can clearly display the period decomposition structure and risk factor decomposition structure of antenatal examination data, so as to facilitate rapid identification and understanding of key information. The present invention can systematically evaluate the risks that pregnant women may face in the whole antenatal examination process through antenatal examination risk tree analysis method, which helps doctors to fully understand the health status and potential problems of pregnant women, so as to formulate more accurate and effective treatment and management plans, and the antenatal examination risk tree risk model intuitively displays various risk factors and their relationships with each other in a tree structure, which can clearly see which factors may have the greatest impact on the health of pregnant women and fetuses, and calculates the probability of occurrence of top events (i.e., adverse pregnancy outcomes) through probability calculation method, so as to provide doctors with quantitative risk assessment results, and can judge the severity of risks according to specific values, so as to make more scientific decisions, which is helpful to improve the health level of mothers and infants.

[0079] It should be understood by those skilled in the art that the embodiments of the present invention may be provided as methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, optical storage, etc.) containing computer-usable program code.

[0080] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for evaluating and analyzing pregnant women in obstetrics and gynecology based on community prenatal examination data stream, characterized in that: The obstetrics and gynecology pregnant woman assessment and analysis method based on community prenatal examination data stream includes the following steps: S1. Build a community prenatal examination data stream storage platform to obtain and store prenatal examination data in real time; S2. Based on the acquired prenatal examination data, a coupling matrix is ​​constructed by decomposing the prenatal examination cycle and risk factors; S3. Based on the constructed coupling matrix, a prenatal examination risk tree risk model is established using the prenatal examination risk tree analysis method, and the prenatal examination risk tree risk model is used to conduct risk assessment on the prenatal examination results; S4. Display risk assessment results through data visualization technology; The step of constructing a coupling matrix based on the acquired prenatal examination data by decomposing the prenatal examination cycle and the risk factors includes the following steps: S21, preprocessing the acquired prenatal examination data, wherein the preprocessing includes data cleaning and standardization; S22, constructing a prenatal examination data classification model, and classifying the prenatal examination data after prenatal examination through the prenatal examination data classification model; S23, performing time series analysis on each type of data in the prenatal examination data, and according to the analysis results, performing period decomposition on the prenatal examination data using a preset period standard; S24. Construct a cycle decomposition structure according to the cycle decomposition results and classification results of the prenatal examination data; S25. According to the inherent attributes of each type of data in the prenatal examination data, identify the risk factors of each type of data, and construct a risk factor decomposition structure through step-by-step decomposition and analysis; S26. Establish a coupling matrix based on the cycle decomposition structure and risk factor decomposition structure.

2. The method for evaluating and analyzing obstetrics and gynecology pregnant women based on community prenatal examination data stream according to claim 1 is characterized in that: The method of constructing a prenatal examination data classification model and classifying the prenatal examination data after prenatal examination by using the prenatal examination data classification model comprises the following steps: S221, collect historical prenatal examination sample data, and establish a training data set through feature selection processing; S222. Construct and train a prenatal examination data classification model using a support vector machine algorithm based on the obtained training data set; S223. The preprocessed antenatal examination data is used as input and input into the trained antenatal examination data classification model to classify the preprocessed antenatal examination data.

3. The method for evaluating and analyzing obstetrics and gynecology pregnant women based on community prenatal examination data stream according to claim 2 is characterized in that: Collecting historical prenatal examination sample data and establishing a training data set through feature selection processing includes the following steps: S2211. Perform dimensionality reduction processing on historical prenatal examination sample data by using dimensionality reduction technology; S2212, performing correlation feature selection on the historical prenatal examination sample data after dimension reduction to establish a feature data set; S2213. Optimize the feature data set using a spectral clustering algorithm to obtain a new feature data set, and use the new feature data set as a training data set.

4. The method for evaluating and analyzing obstetrics and gynecology pregnant women based on community prenatal examination data stream according to claim 3 is characterized in that: The method of processing the feature data set by using the spectral clustering algorithm to obtain a new feature data set includes the following steps: S22131. For each feature data point in the feature data set, calculate the similarity between each pair of feature data points; S22132, constructing an undirected graph using each feature data point in the feature data set as a node, and using the similarity between the feature data points as the weight of the edge between the nodes in the undirected graph; S22133, constructing a weighted connection matrix and a degree matrix according to the similarity between nodes and the sum of the weights of all edges connected to each node; S22134. Calculate a normalized Laplace matrix based on the constructed weighted connection matrix and degree matrix; S22135, performing eigenvalue decomposition on the calculated normalized Laplace matrix, and extracting eigenvectors corresponding to several minimum non-zero eigenvalues ​​therein; S22136. Divide the data points into several clusters according to the feature vectors, and select the cluster center from each cluster as a representative point to form a new feature data set.

5. The method for evaluating and analyzing obstetrics and gynecology pregnant women based on community prenatal examination data stream according to claim 4 is characterized in that: The calculation formula for calculating the standardized Laplace matrix based on the constructed weighted connection matrix and degree matrix is: R=F -1 / 2 W F -1 / 2 ; Where R represents the normalized Laplace matrix; F represents the degree matrix; F -1 / 2 represents the diagonal matrix formed by taking the square root of the reciprocal of each diagonal element of the degree matrix F; W represents the weighted connection matrix.

6. The method for evaluating and analyzing obstetrics and gynecology pregnant women based on community prenatal examination data stream according to claim 1, characterized in that: The step of performing time series analysis on each type of data in the prenatal examination data and periodically decomposing the prenatal examination data according to the analysis results by using a preset period standard comprises the following steps: S231, obtaining the real-time acquisition time of the prenatal examination data, and arranging each type of data in the prenatal examination data in the order of the real-time acquisition time to form a time series; S232, dividing each type of data in the prenatal examination data into periods according to a preset period standard; S233, grouping and arranging the prenatal examination data after the cycle division according to the cycle.

7. The obstetrics and gynecology pregnant woman assessment and analysis method based on community prenatal examination data stream according to claim 1 is characterized in that: The method of establishing a prenatal examination risk tree risk model based on the constructed coupling matrix and using the prenatal examination risk tree analysis method, and conducting risk assessment on the prenatal examination results through the prenatal examination risk tree risk model includes the following steps: S31, determine the top event of the prenatal examination risk tree, use the coupling matrix to analyze the top event of the prenatal examination risk tree, identify the basic events and build a prenatal examination risk tree risk model; S32. Calculate the probability of occurrence of the top event by probability calculation method according to the antenatal examination risk tree risk model; S33. Calculate the sensitivity coefficient of each basic event through sensitivity analysis according to the probability of occurrence of the top event; S34. Conduct risk assessment on prenatal examination results based on sensitivity analysis results.

8. The method for evaluating and analyzing obstetrics and gynecology pregnant women based on community prenatal examination data stream according to claim 7, characterized in that: The method of calculating the probability of occurrence of the top event by a probability calculation method according to the prenatal examination risk tree risk model includes the following steps: S321, assigning a basic probability value to each basic event in the antenatal examination risk tree risk model; S322, determining the logical relationship between events according to the logic gates in the antenatal examination risk tree risk model; S323. Starting from the basic event at the bottom of the antenatal examination risk tree risk model, calculate upward step by step according to the logical relationship to obtain the probability of the top event occurring.

9. The method for evaluating and analyzing obstetrics and gynecology pregnant women based on community prenatal examination data stream according to claim 8, characterized in that: The calculation formula for calculating the sensitivity coefficient of each basic event through sensitivity analysis based on the probability of occurrence of the top event is: ； In the formula, H i Represents basic events i The sensitivity coefficient of Indicates top events T The partial derivative value of the probability value of occurrence; D ( T ) indicates the top event T Probability value of occurrence; Represents basic events i The partial derivative value of the basic probability value of ; k i Represents basic events i The basic probability value of .

10. A system for evaluating and analyzing obstetric and gynecological pregnant women based on community prenatal examination data stream, used to implement the method for evaluating and analyzing obstetric and gynecological pregnant women based on community prenatal examination data stream according to any one of claims 1 to 9, characterized in that: The obstetrics and gynecology pregnant woman assessment and analysis system based on community prenatal examination data stream includes: a data existence module, a matrix construction module, a risk assessment module and a result display module, wherein the data existence module, the matrix construction module, the risk assessment module and the result display module are connected in sequence; The data storage module is used to build a community prenatal examination data stream storage platform to obtain and store prenatal examination data in real time; The matrix construction module is used to construct a coupling matrix through prenatal examination cycle decomposition and risk factor decomposition based on the acquired prenatal examination data; The risk assessment module is used to establish a prenatal examination risk tree risk model based on the constructed coupling matrix using the prenatal examination risk tree analysis method, and to perform risk assessment on the prenatal examination results through the prenatal examination risk tree risk model; The result display module is used to display the risk assessment results through data visualization technology.

Images