Insurance recommendation evaluation method and device applicable to assisted reproductive insurance

Abstract

The present invention provides an insurance recommendation evaluation method and device applicable to assisted reproductive insurance, comprising: receiving index data input by an applicant for reproductive insurance, wherein the index data includes any one or more of physical examination result information and physiological characteristic information; processing the index data based on a preset prediction model to generate an insurance recommendation result, wherein the prediction model is pre-trained and generated. The technical solution provided by the present invention can accurately and quickly confirm the insurance eligibility of a customer, provide risk assessment for assisted reproductive insurance underwriting, make the insurance company's pricing more accurate, and provide improved risk classification for insurance decision support and risk management, thereby providing high-quality health management and protection services for the insured.

Description

Technical Field

[0001] The present invention relates to artificial intelligence technology, and in particular to an insurance recommendation evaluation method and device suitable for assisted reproductive insurance. Background Art

[0002] Existing assisted reproductive insurance cooperates with specific assisted reproductive medical centers, and then the reproductive medical centers evaluate the overall risk attributes of assisted reproductive insurance to decide whether to provide protection for individual policyholders. Before the specific assisted reproductive medical center conducts the evaluation, the insurance company cannot timely and accurately identify the risk attributes of individual policyholders, and thus cannot provide pricing advantages, nor can it cooperate with more medical institutions and provide protection services for more assisted reproductive patients. Summary of the invention

[0003] The embodiments of the present invention provide an insurance recommendation evaluation method and device suitable for assisted reproductive insurance, which can accurately and quickly confirm the customer's insurance eligibility, provide risk assessment for assisted reproductive insurance underwriting, make the insurance company's pricing more accurate, and provide improved risk classification for insurance decision support and risk management, thereby providing high-quality health management and protection services for policyholders.

[0004] A first aspect of an embodiment of the present invention provides an insurance recommendation evaluation method applicable to assisted reproductive insurance, comprising:

[0005] Receiving index data input by an applicant for reproductive insurance coverage, the index data including any one or more of physical examination result information and physiological characteristic information;

[0006] The indicator data is processed based on a preset prediction model to generate an insurance recommendation result, wherein the prediction model is pre-trained and generated.

[0007] Optionally, in a possible implementation of the first aspect, the indicator data is processed based on a preset prediction model to generate an insurance recommendation result, wherein the prediction model is pre-trained and generated including:

[0008] Obtain a preset number of samples from the database as a training sample set;

[0009] Based on the preset algorithm, the number of pregnancies of each couple through assisted reproduction is used as the prediction target, and the original variables or derived variables after cleaning the training sample set are used as the training feature set to obtain the correlation function between the prediction target and the training feature set;

[0010] The correlation function between the prediction target and the training feature set is optimized to obtain a final parameter set.

[0011] Optionally, in a possible implementation manner of the first aspect, the prediction model is pre-trained and generated based on a stratified random sampling method.

[0012] Optionally, in a possible implementation of the first aspect, the optimal hyperparameters of the prediction model are determined based on a K-fold cross-validation grid search to obtain a hyperparameter set that maximizes the average accuracy;

[0013] The prediction models are fitted to the entire training data set respectively.

[0014] Optionally, in a possible implementation manner of the first aspect, the preset algorithm is any one or more of a logistic regression algorithm, a random forest algorithm, and a catboost algorithm.

[0015] A second aspect of an embodiment of the present invention provides an insurance recommendation evaluation device applicable to assisted reproductive insurance, comprising:

[0016] An applicant data receiving module, used to receive index data input by an applicant for reproductive insurance, wherein the index data includes any one or more of physical examination result information and physiological characteristic information;

[0017] The insurance recommendation result generating module is used to process the indicator data based on a preset prediction model to generate an insurance recommendation result, wherein the prediction model is pre-trained and generated.

[0018] Optionally, in a possible implementation manner of the second aspect, the indicator data is processed based on a preset prediction model to generate an insurance recommendation result, wherein the prediction model is pre-trained and generated including:

[0019] Obtain a preset number of samples from the database as a training sample set;

[0020] Based on the preset algorithm, the number of pregnancies of each couple through assisted reproduction is used as the prediction target, and the original variables or derived variables after cleaning the training sample set are used as the training feature set to obtain the correlation function between the prediction target and the training feature set;

[0021] The correlation function between the prediction target and the training feature set is optimized to obtain a final parameter set.

[0022] Optionally, in a possible implementation manner of the second aspect, the prediction model is pre-trained and generated based on a stratified random sampling method.

[0023] Optionally, in a possible implementation of the second aspect, the optimal hyperparameters of the prediction model are determined based on a K-fold cross-validation grid search to obtain a hyperparameter set that maximizes the average accuracy;

[0024] The prediction models are fitted to the entire training data set respectively.

[0025] The third aspect of the embodiment of the present invention provides a readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, it is used to implement the method of the first aspect of the present invention and various possible designs of the first aspect.

[0026] The present invention provides a method and device for evaluating insurance recommendations for assisted reproductive insurance. The intelligent underwriting engine, which is based on medical health data and has an artificial intelligence algorithm as its core, is the first development and application of assisted reproductive insurance. The intelligent underwriting engine evaluates insurance recommendations through a predictive model.

[0027] The present invention can accurately and quickly confirm the insurance eligibility of customers, provide risk assessment for assisted reproductive insurance underwriting, make insurance companies' pricing more accurate, and provide improved risk classification for insurance decision support and risk management, thereby providing high-quality health management and protection services for policyholders. This underwriting engine screens based on personal risk attributes, and the application process is efficient and fast. It can conduct business in a wide range of assisted reproductive institutions, thereby providing more patients with high-quality, continuous and stable protection financial services. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 A flowchart of a first implementation method for proposing an assessment method for insurance coverage applicable to assisted reproduction;

[0029] Figure 2 A flowchart of a second implementation method for proposing an assessment method for assisted reproductive insurance;

[0030] Figure 3 Flowchart of risk assessment for assisted reproduction insurance;.

[0031] Figure 4 A structural diagram of the assessment device for the proposed insurance coverage of assisted reproduction; DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0033] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in sequences other than those illustrated or described herein.

[0034] It should be understood that in various embodiments of the present invention, the size of the sequence number of each process does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0035] It should be understood that in the present invention, "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products or apparatuses.

[0036] It should be understood that in the present invention, "plurality" refers to two or more than two. "And / or" is merely a description of the association relationship of associated objects, indicating that three relationships may exist. For example, and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "Contains A, B and C", "Contains A, B, C" means that A, B, and C are all included, "Contains A, B or C" means that one of A, B, and C is included, and "Contains A, B and / or C" means that any one, any two, or any three of A, B, and C are included.

[0037] It should be understood that in the present invention, "B corresponding to A", "B corresponding to A", "A corresponds to B" or "B corresponds to A" means that B is associated with A and B can be determined based on A. Determining B based on A does not mean determining B based only on A, but B can also be determined based on A and / or other information. A and B match when the similarity between A and B is greater than or equal to a preset threshold.

[0038] Depending on the context, "if" as used herein may be interpreted as "when" or "when" or "in response to determining" or "in response to detecting."

[0039] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0040] The present invention provides an insurance recommendation evaluation method applicable to assisted reproductive insurance, such as Figure 1 The flow chart shown includes:

[0041] S10, receiving the index data input by the applicant for reproductive insurance, wherein the index data includes any one or more of the physical examination result information and the physiological characteristic information, wherein the physical examination result information and the physiological characteristic information respectively include age, ovarian reserve (AMH, AFC, FSH, etc.), body mass index BMI, six hormones, semen analysis, duration of infertility, previous live births, previous miscarriages and infertility type, etc.

[0042] S20. Process the indicator data based on a preset prediction model to generate an insurance recommendation result, wherein the prediction model is pre-trained and generated.

[0043] In one embodiment, Figure 2 As shown, the indicator data is processed based on a preset prediction model to generate an insurance recommendation result, wherein the prediction model is pre-trained and generated including:

[0044] S201, obtaining a preset number of samples in a database as a training sample set;

[0045] S202, based on a preset algorithm, taking the number of pregnancies of each couple through assisted reproduction as a prediction target, and the original variables or derived variables after cleaning the training sample set as a training feature set, to obtain a correlation function between the prediction target and the training feature set;

[0046] S203: Optimize the correlation function between the prediction target and the training feature set to obtain a final parameter set.

[0047] In the above steps, after data processing, 70% of the data set is randomly selected as a training set to build a prediction model, and the remaining 30% is used for validation. A variety of supervised machine learning algorithms are used to construct the main components of the prediction model / supplemented model to facilitate the subsequent reference of logistic regression and stratified random sampling, including logistic regression, random forest, and catboost. Although the above three algorithms have their own advantages and disadvantages, they all have interpretable properties. The calculation process can be repeated ("simulatability") with a full understanding of the algorithm, and each part of the model has an intuitive explanation ("algorithm transparency"). The prediction model takes the number of assisted reproductive cycles each couple needs to get pregnant as the prediction target, and the cleaned original variables or derived variables are used as the training feature set. The interaction relationship between the training features is mined, and the final parameter set is obtained by optimizing the objective function.

[0048] Furthermore, the prediction model is pre-trained and generated based on a stratified random sampling method.

[0049] Furthermore, the optimal hyperparameters of the prediction model are determined based on a grid search with K-fold cross-validation to obtain the hyperparameter set that maximizes the average accuracy;

[0050] The prediction models are fitted to the entire training data set respectively.

[0051] The training process of the prediction model adopts a stratified random sampling method to ensure that the proportion of live births and non-live births in the training set and the validation set is the same as that in the original data set. The optimal hyperparameters of the algorithm are selected by a grid search with K-fold cross-validation, and finally the hyperparameter set with the maximum average accuracy is obtained, and the model is fitted to the entire training data set respectively. Finally, the present invention evaluates the receiver operating characteristic (ROC) curve to select the final model.

[0052] Since the present invention is used for risk assessment of assisted reproductive insurance before entering the cycle, such as Figure 3 As shown, the present invention can therefore only use clinical data obtained before starting the IVF procedure.

[0053] The present invention trains and develops artificial intelligence algorithms based on a large amount of health physiological and clinical data, and verifies the algorithms with reserved data other than training data. The risk attributes of applicants for assisted reproductive insurance are evaluated when the applicants provide some physical examination results and physiological characteristics. By identifying adverse events covered by insurance and speeding up the application and underwriting process, effective solutions are provided for insurance companies in the fields of insurance consulting, application process, underwriting process and risk management, enabling them to provide customers with high-quality and affordable assisted reproductive protection and health management services.

[0054] In the initial data for model training, anonymous data of more than 100,000 couples recorded by the Human Fertilization and Embryology Authority (HFEA), which is responsible for regulating fertility treatment in the UK, was used. Anonymous data is the registration data that includes most of the data since 1991 after the patient's identity information is hidden. Anonymous data without identification codes protects patient privacy by grading or hiding some information. Scientific research using anonymous data can still answer important questions. Such as which factors (patients, types of treatments, and technologies used) can affect treatment outcomes, or which indicators suggest that patients only need to implant one or two embryos during treatment. This information covers cycles from January 1, 1991 to March 31, 2009. This data set contains about 90 variables. Not all variables have a significant predictive effect on live births. Under the guidance of reproductive medical experts, the present invention cleans and corrects the data again and clearly defines and standardizes the variables to reproduce candidate predictive variables.

[0055] In addition, contradictory data points in the original data are eliminated, such as the use of frozen embryos in the later stage without a previous history of pregnancy and childbirth; the original variables are used to generate derived variables, such as weight and height to calculate the body mass index (BMI), and classify and label them; the causes of infertility and reproductive medical experts are classified and scored, etc.

[0056] In the above initial training and verification process, by evaluating the importance of parameter features to the final prediction results, multiple relevant variables of both men and women before entering the assisted reproductive week were extracted, such as age, ovarian reserve (AMH, AFC, FSH, etc.), body mass index BMI, six hormones, semen analysis, duration of infertility, previous live births, previous miscarriages and infertility types, etc., and the applicants need to do several IVF cycles for successful pregnancy to predict. Combine the prediction results and perform feature importance ranking through retrospective analysis to find the most relevant features of the applicants to finally evaluate the risk attributes of the applicants for assisted reproductive insurance: extremely high risk, high risk, moderate risk, medium-low risk and low risk. / Classification name to add risk attributes /

[0057] After determining the final prediction of the risk assessment of assisted reproductive insurance, the final first-generation underwriting engine is obtained by fitting the entire training data. In addition, it participates in the domestic assisted reproductive "good pregnancy insurance" insurance business, and uses data from multiple reproductive medical centers for verification and iteration.

[0058] A second aspect of an embodiment of the present invention provides an insurance recommendation evaluation device applicable to assisted reproductive insurance, such as Figure 4 As shown, including:

[0059] An applicant data receiving module, used to receive index data input by an applicant for reproductive insurance, wherein the index data includes any one or more of physical examination result information and physiological characteristic information;

[0060] The insurance recommendation result generating module is used to process the indicator data based on a preset prediction model to generate an insurance recommendation result, wherein the prediction model is pre-trained and generated.

[0061] Furthermore, the indicator data is processed based on a preset prediction model to generate an insurance recommendation result, wherein the prediction model is pre-trained and generated including:

[0062] Obtain a preset number of samples from the database as a training sample set;

[0063] Based on the preset algorithm, the number of pregnancies of each couple through assisted reproduction is used as the prediction target, and the original variables or derived variables after cleaning the training sample set are used as the training feature set to obtain the correlation function between the prediction target and the training feature set;

[0064] The correlation function between the prediction target and the training feature set is optimized to obtain a final parameter set.

[0065] Furthermore, the prediction model is pre-trained and generated based on a stratified random sampling method.

[0066] Furthermore, the optimal hyperparameters of the prediction model are determined based on a grid search with K-fold cross-validation to obtain the hyperparameter set that maximizes the average accuracy;

[0067] The prediction models are fitted to the entire training data set respectively.

[0068] Among them, the readable storage medium can be a computer storage medium or a communication medium. The communication medium includes any medium that facilitates the transmission of a computer program from one place to another. The computer storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer. For example, a readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in an application-specific integrated circuit (Application Specific Integrated Circuits, abbreviated as: ASIC). In addition, the ASIC can be located in a user device. Of course, the processor and the readable storage medium can also exist in a communication device as discrete components. The readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0069] The present invention also provides a program product, which includes an execution instruction, which is stored in a readable storage medium. At least one processor of a device can read the execution instruction from the readable storage medium, and at least one processor executes the execution instruction so that the device implements the methods provided in the above various embodiments.

[0070] In the above-mentioned terminal or server embodiments, it should be understood that the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the present invention may be directly implemented as being executed by a hardware processor, or may be executed by a combination of hardware and software modules in the processor.

[0071] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for evaluating insurance recommendations for assisted reproductive insurance, characterized in that: include: Receiving indicator data input by an applicant for reproductive insurance coverage, wherein the indicator data includes any one or more of physical examination result information and physiological characteristic information, and the indicator data is clinical data obtained before starting the IVF procedure; Processing the indicator data based on a preset prediction model to generate an insurance recommendation result, wherein the prediction model is pre-trained and generated; The indicator data is processed based on a preset prediction model to generate an insurance recommendation result, wherein the prediction model is pre-trained and generated including: Obtain a preset number of samples from the database as a training sample set; Based on a preset algorithm, the number of pregnancies each couple experienced through assisted reproduction was used as the prediction target, and the original variables or derived variables after cleaning the training sample set were used as the training feature set to obtain the correlation function between the prediction target and the training feature set. Optimizing the correlation function between the prediction target and the training feature set to obtain a final parameter set; The prediction model is pre-trained and generated based on a stratified random sampling method; The optimal hyperparameters of the prediction model are determined based on a grid search using K-fold cross-validation, and the hyperparameter set that maximizes the average accuracy is obtained. Fitting the prediction model to the entire training data set respectively; The preset algorithm is any one or more of a logistic regression algorithm, a random forest algorithm, and a catboost algorithm; The method further comprises: Combine the prediction results and perform feature importance ranking through retrospective analysis to find the most relevant features of the applicants and ultimately assess the risk attributes of applicants for assisted reproductive insurance. After determining the final prediction of the risk assessment for assisted reproductive insurance, a final fit is performed on the entire training data to produce the final first-generation underwriting engine; Data from multiple reproductive medical centers were used for validation and iteration.

2. A device for evaluating insurance recommendations for assisted reproductive insurance, characterized in that: include: An applicant data receiving module, configured to receive index data input by an applicant for reproductive insurance, wherein the index data includes any one or more of physical examination result information and physiological characteristic information; An insurance recommendation result generating module, configured to process the indicator data based on a preset prediction model to generate an insurance recommendation result, wherein the prediction model is pre-trained and generated; The indicator data is processed based on a preset prediction model to generate an insurance recommendation result, wherein the prediction model is pre-trained and generated including: Obtain a preset number of samples from the database as a training sample set; Based on a preset algorithm, the number of pregnancies each couple experienced through assisted reproduction was used as the prediction target, and the original variables or derived variables after cleaning the training sample set were used as the training feature set to obtain the correlation function between the prediction target and the training feature set. Optimizing the correlation function between the prediction target and the training feature set to obtain a final parameter set; The prediction model is pre-trained and generated based on a stratified random sampling method; The optimal hyperparameters of the prediction model are determined based on a grid search using K-fold cross-validation, and the hyperparameter set that maximizes the average accuracy is obtained. Fitting the prediction model to the entire training data set respectively; The preset algorithm is any one or more of a logistic regression algorithm, a random forest algorithm, and a catboost algorithm; The device is also used for: Combine the prediction results and perform feature importance ranking through retrospective analysis to find the most relevant features of the applicants and ultimately assess the risk attributes of applicants for assisted reproductive insurance. After determining the final prediction of the risk assessment for assisted reproductive insurance, a final fit is performed on the entire training data to produce the final first-generation underwriting engine; Data from multiple reproductive medical centers were used for validation and iteration.

3. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which is used to implement the method according to claim 1 when executed by a processor.

Images