Quality control method, device and equipment for eye health screening data and medium

Abstract

The invention discloses a quality control method and device for eye health screening data, equipment and a medium. The method comprises the steps of obtaining initial eye health screening data of a to-be-detected target; carrying out integrity verification and accuracy verification on the basic identity information; performing index accuracy verification on eye health index data in the first eye health screening data; obtaining each index value in the second eye health screening data, and determining to-be-judged eye health index data according to each index value and a preset index relationship; correcting the to-be-judged eye health index data which do not pass the judgment; and inputting the third eye health screening data into the pre-training prediction model to screen the eye health screening data of the to-be-detected target. Error data can be prevented from entering the central database, and the purity and consistency of the data are further guaranteed. Data errors are controlled in a screening site, diffusion and amplification of error data are avoided, and accuracy and reliability of the data are guaranteed.

Description

Technical Field

[0001] The present application relates to the field of eye health data quality inspection, and specifically to a quality control method, device, equipment and medium for eye health screening data. Background Art

[0002] The vision of children and adolescents is screened regularly, and errors in the screening process are often difficult to detect and correct immediately at the screening site. On the one hand, screeners usually need to complete the screening task within a limited time, and it is difficult to have sufficient time to check and verify the data one by one. On the other hand, even if they find errors, they may not be able to correct them immediately due to equipment and site limitations. This has seriously affected the accuracy and reliability of the data. Even if these errors are discovered through data comparison and review afterwards, it is often difficult to make perfect corrections due to the huge amount of data and tight time. This not only increases the difficulty and cost of data processing, but may also affect subsequent analysis and decision-making. Therefore, in order to ensure the accuracy and reliability of school screening data, we need to establish an effective data quality inspection and correction mechanism so that errors can be corrected and supplemented in a timely manner after they are discovered. Summary of the invention

[0004] In order to solve the above problems, the present application proposes a quality control method, device and medium for eye health screening data, wherein the method includes:

[0005] Obtain initial eye health screening data of the target to be measured, wherein the initial eye health screening data includes basic identity information and eye health indicator data; perform integrity check and accuracy check on the basic identity information to obtain first eye health screening data; perform indicator accuracy check on the eye health indicator data in the first eye health screening data to obtain second eye health screening data; obtain the value of each indicator in the second eye health screening data, and determine the eye health indicator data to be determined through the values ​​of each indicator and the preset indicator relationship; correct the eye health indicator data to be determined that fails the determination to obtain third eye health screening data; input the third eye health screening data into a pre-trained prediction model to screen the eye health screening data of the target to be measured.

[0006] In one example, the basic identity information includes at least name, gender, date of birth, ID number, school, grade, class, contact number, and contact address; and / or, the integrity check and accuracy check of the basic identity information specifically include: determining the required information in the basic identity information, and determining whether the required information is complete; using regular expressions to perform format checks on the ID number and telephone number; using an address library to check the contact address, and checking the date format of the date of birth; determining the data source of the basic identity information, and if the basic identity information has multiple data sources, performing consistency checks on the basic identity information of the multiple data sources.

[0007] In one example, the eye health indicator data in the first eye health screening data include: eye axis, visual acuity, refractive status, cornea, pupil distance, intraocular pressure, accommodation amplitude, accommodation sensitivity, stereoscopic vision, amount of latent strabismus, ratio of accommodative convergence to accommodation, corneal diameter, Worth's four points, convergence near point breaking point, blur point, recovery point, positive and negative flip shooting through difficult lenses, long-distance BI examination, long-distance BO examination, near BI examination and near BO examination.

[0008] In one example, obtaining the values ​​of each indicator in the second eye health screening data and determining the eye health indicator data to be determined through the values ​​of each indicator and the preset indicator relationship specifically includes: determining a first eye health indicator and a second eye health indicator in the second eye health screening data; the first eye health indicator is an eye health indicator that can be used to predict the eye state, and the second eye health indicator is an eye health indicator that can be obtained by predicting the eye state; determining the eye state prediction value of the target to be measured through the first eye health indicator; determining the second eye health indicator value prediction interval of the target to be measured based on the eye state prediction value; if the indicator value corresponding to the second eye health indicator is not within the second eye health indicator value prediction interval, the indicator value corresponding to the second eye health indicator is the eye health indicator data to be determined.

[0009] In one example, the first eye health indicator includes at least one of the eye axis and the equivalent spherical diopter after mydriasis; the eye state includes at least one of myopia and hyperopia; the second eye health indicator includes at least one of the equivalent spherical diopter after mydriasis, spherical lens, and naked eye visual acuity.

[0010] In one example, the third eye health screening data is input into a pre-trained prediction model to screen the eye health screening data of the target to be tested, specifically including: selecting part of the data in the third eye health screening data set as a prediction set, and using data other than the prediction set as a comparison set; the part of the data includes basic identity information and / or eye health indicator data; the prediction set is input into a preset prediction model to obtain a prediction result; the prediction result is compared with the comparison set to obtain a comparison error; if the comparison error is greater than a preset threshold, the data corresponding to the comparison set is screened.

[0011] In one example, before inputting the prediction set into the preset prediction model, the method also includes: obtaining historical eye health indicator data of eye health indicators; in addition to basic identity information and eye health indicator data, the historical eye health indicator data also includes the age of myopia onset, parents' educational level, parents' myopia status, and outdoor exercise time obtained from questionnaire survey information; performing principal component analysis on each indicator to obtain characteristic indicators, and determining the characteristic data corresponding to the characteristic indicators in the historical eye health indicator data; after deleting the missing values ​​in the characteristic data, dividing the characteristic data into a training set and a test set; using the training set to train the pre-trained prediction model, and evaluating the model performance through cross-validation.

[0012] The present application also provides a quality control device for eye health screening data, comprising: a data acquisition module, which acquires initial eye health screening data of a target to be measured, wherein the initial eye health screening data includes basic identity information and eye health indicator data; a first screening module, which performs integrity verification and accuracy verification on the basic identity information to obtain first eye health screening data; a second screening module, which performs indicator accuracy verification on the eye health indicator data in the first eye health screening data to obtain second eye health screening data; an indicator connection module, which acquires the value of each indicator in the second eye health screening data, and determines the eye health indicator data to be determined through the value of each indicator and the preset indicator relationship; a third screening module, which corrects the eye health indicator data to be determined that has not passed the determination to obtain third eye health screening data; a prediction screening module, which inputs the third eye health screening data into a pre-trained prediction model to screen the eye health screening data of the target to be measured.

[0013] The present application also provides a quality control device for eye health screening data, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute: obtaining initial eye health screening data of a target to be measured, the initial eye health screening data comprising basic identity information and eye health indicator data; performing integrity check and accuracy check on the basic identity information to obtain first eye health screening data; performing indicator accuracy check on the eye health indicator data in the first eye health screening data to obtain second eye health screening data; obtaining the value of each indicator in the second eye health screening data, and determining the eye health indicator data to be determined through the values ​​of each indicator and the preset indicator relationship; correcting the eye health indicator data to be determined that has not passed the determination to obtain third eye health screening data; and inputting the third eye health screening data into a pre-trained prediction model to screen the eye health screening data of the target to be measured.

[0014] The present application also provides a non-volatile computer storage medium storing computer executable instructions, wherein the computer executable instructions are configured to: obtain initial eye health screening data of a target to be measured, wherein the initial eye health screening data includes basic identity information and eye health indicator data; perform integrity check and accuracy check on the basic identity information to obtain first eye health screening data; perform indicator accuracy check on the eye health indicator data in the first eye health screening data to obtain second eye health screening data; obtain a value of each indicator in the second eye health screening data, and determine the eye health indicator data to be determined through the values ​​of each indicator and a preset indicator relationship; correct the eye health indicator data to be determined that has not passed the determination to obtain third eye health screening data; and input the third eye health screening data into a pre-trained prediction model to screen the eye health screening data of the target to be measured.

[0015] The method proposed in this application can bring the following beneficial effects: through three-level real-time audit, the eye health screening data is rigorously and meticulously controlled at the data collection site, thereby ensuring the accuracy and completeness of the data. The three-level real-time audit can prevent the occurrence of data errors from the source, and can immediately conduct on-site confirmation and rectification when errors are found, thereby avoiding the adverse effects of erroneous data on subsequent analysis and decision-making. In this way, erroneous data can be prevented from entering the central database, further ensuring the purity and consistency of the data. Data errors are controlled at the screening site to avoid the spread and amplification of erroneous data. The accuracy and reliability of the data are guaranteed, and at the same time a solid and reliable basis is provided for subsequent data analysis and decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0017] Figure 1 This is a flow chart of a quality control method for eye health screening data in an embodiment of the present application;

[0018] Figure 2 This is a schematic diagram of the association of eye health indicators in a quality control method for eye health screening data in an embodiment of the present application;

[0019] Figure 3 This is a schematic diagram of the structure of a quality control device for eye health screening data in an embodiment of the present application;

[0020] Figure 4 This is a schematic diagram of the structure of a quality control device for eye health screening data in an embodiment of the present application. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0022] Myopia is a common and widespread eye disease, most common among children and adolescents. The detection rate of myopia among urban and rural students continues to rise, and there is also a clear trend of younger age. Regularly screening the vision of children and adolescents, detecting vision abnormalities early and correcting them is of great significance to ensuring the vision health of children and adolescents.

[0023] However, in the actual operation of vision screening for school admission, we are currently faced with huge data collection challenges and strict time constraints. First, in terms of personnel, screeners often need to complete the vision examinations of a large number of students in a short period of time, which may cause them to become fatigued or negligent during the operation. For example, they may neglect to calibrate the equipment in a hurry to complete the task, or make typos or misreadings when entering data. In addition, if the screeners lack sufficient training or experience, they may also be unable to make correct judgments and handle special situations. Secondly, in terms of the data collection information system, if its design is unreasonable or there are loopholes, it may cause data to be lost or deformed during transmission or storage. In addition, since the environment at the screening site is usually noisy, the equipment may be interfered with by various external factors, such as unstable power supply, temperature changes, electromagnetic interference, etc., resulting in abnormal operation of the equipment or inaccurate data reading. In addition, if the screening equipment is used for a long time without proper maintenance, it may also wear out or malfunction, thereby increasing the risk of data errors.

[0024] What is even more difficult is that these errors are often difficult to detect and correct immediately at the screening site. On the one hand, screeners usually need to complete the screening task within a limited time, and it is difficult to have sufficient time to check and verify the data one by one. On the other hand, even if they find errors, they may not be able to correct them immediately due to equipment and site limitations. This has seriously affected the accuracy and reliability of the data. Even if these errors are discovered through data comparison and review afterwards, it is often difficult to make perfect corrections due to the huge amount of data and tight time. This not only increases the difficulty and cost of data processing, but may also affect subsequent analysis and decision-making.

[0025] Therefore, in order to ensure the accuracy and reliability of school screening data, we need to establish an effective data quality inspection and correction mechanism so that errors can be corrected and supplemented in a timely manner after they are discovered.

[0026] At present, the control of eye health screening data quality is mainly reflected in the quality review of the inspection site and the data cleaning and standardization when uploading to the database. The quality control of the inspection site refers to the random sampling of 5% of the retest subjects for vision retest and error inspection during the daily on-site investigation. When the test error exceeds 5%, the on-site discussion with the physical examination physician is carried out until the test quality is marked before continuing the test; if the incidence rate of the number of indicators with errors is greater than 10%, all vision test data on that day will be deemed invalid and must be retested, otherwise the data measured on that day cannot be included in the statistics. This quality control method provides a guarantee for the improvement of data quality to a certain extent, but it consumes more time and energy. And it cannot fully guarantee the validity of every piece of data.

[0027] Data cleaning and standardization when uploading to the database refers to the use of ETL tools to clean the data. This method ensures the quality of data uploaded to the data warehouse, but it cannot control the quality of data at the screening site, which will cause a large amount of screening data to be invalidated and waste manpower and material resources.

[0028] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0029] Figure 1 A flowchart of a quality control method for eye health screening data provided by one or more embodiments of this specification. The method can be applied to the entry of eye health data of students of different ages. The process can be executed by a computing device in the corresponding field. Certain input parameters or results in the process allow manual intervention and adjustment to help improve accuracy.

[0030] The analysis method involved in the embodiments of the present application can be implemented by a terminal device or a server, and the present application does not impose any special restrictions on this. For the convenience of understanding and description, the following embodiments are described in detail by taking a server as an example.

[0031] It should be noted that the server can be a single device or a system composed of multiple devices, that is, a distributed server, and this application does not make any specific limitations on this.

[0032] like Figure 1 As shown, the embodiment of the present application provides a method for quality control of eye health screening data, comprising:

[0033] S101: Acquire initial eye health screening data of a target to be tested, wherein the initial eye health screening data includes basic identity information and eye health indicator data.

[0034] The target to be tested here refers to students who need to enter eye health data, and the initial eye health screening data refers to eye health data that has not been screened. At the collection site, the optometrist uses a data collection terminal (such as a mobile phone APP) to connect various ophthalmic equipment and instruments via Bluetooth. After the connection is successful, the optometry data can be uploaded to the data management background through the data interface or Bluetooth module. After entering the information, the data quality control model will perform the first screening of the data.

[0035] S102: Perform integrity check and accuracy check on the basic identity information to obtain first-eye health screening data.

[0036] During the first screening, the main focus is on checking the integrity and accuracy of the students' basic identity information. Integrity checking means that the required information such as name, gender, ID number, and school have been filled in completely without omissions; accuracy checking means that for key information such as ID number and phone number, regular expressions are used to check the format. The ID number verification also needs to consider the calculation of the area code, birthday code, and check code to ensure that the input information conforms to the specified format; the address information is checked using the address library to ensure the authenticity and accuracy of the address; for date information, check whether it conforms to the date format and ensure that it is within a reasonable range. At the same time, data from multiple sources must be compared and checked to ensure data consistency.

[0037] The rules for verifying students’ basic identity information are shown in the following table:

[0038]

[0039] S103: performing an indicator accuracy check on the eye health indicator data in the first eye health screening data to obtain second eye health screening data.

[0040] After the first screening of the basic identity information, it is necessary to verify the accuracy of the eye health indicator data, that is, to conduct a second screening of the eye health data.

[0041] In one embodiment, when performing the second screening, the professional medical knowledge and experience of medical experts are added, and the professional medical experience includes the judgment of the size and range of the eye health indicators themselves and the mutual judgment between different eye indicators.

[0042] Specifically, the calibration standards for each indicator are as follows:

[0043]

[0044]

[0045]

[0046] S104: Obtain the value of each indicator in the second eye health screening data, and determine the eye health indicator data to be determined through the value of each indicator and the preset indicator relationship.

[0047] Since the eye data are highly correlated, after verifying the eye health indicator data, the connection between the indicators can be established according to whether the student is nearsighted. This connection is only used for judgment during data entry.

[0048] In one embodiment, a first eye health indicator and a second eye health indicator can be determined in the second eye health screening data; the first eye health indicator is an eye health indicator that can be used to predict the eye state, and the second eye health indicator is an eye health indicator that can be obtained by predicting the eye state; through the first eye health indicator, the eye state prediction value of the target to be measured is determined; based on the eye state prediction value, the second eye health indicator value prediction interval of the target to be measured is determined; if the indicator value corresponding to the second eye health indicator is not within the second eye health indicator value prediction interval, the indicator value corresponding to the second eye health indicator is the eye health indicator data to be determined.

[0049] Among them, the first eye health indicator includes at least one of the eye axis and the equivalent spherical diopter after mydriasis; the eye state includes at least one of myopia and hyperopia; the second eye health indicator includes at least one of the equivalent spherical diopter after mydriasis, spherical lens, and naked eye vision.

[0050] Specifically, Figure 2 As shown, a bidirectional arrow indicates that bidirectional judgment can be made, and a unidirectional arrow indicates that only unidirectional judgment can be made. For example, based on the axial length > 24mm, it can be inferred that the student may be myopic, but it cannot be inferred that the axial length of the student may be > 24mm based on myopia. When judging that a student may be myopic based on a certain indicator, it is inferred that other indicators should meet the corresponding range, and based on the inferred range, it is judged whether the submitted data is wrong, and then a reminder window pops up.

[0051] For example, if a student's axial length is >24mm, it is inferred that the student may be myopic, and his naked eye vision may be <5.0, the spherical lens may be <-0.5, and the equivalent spherical lens after mydriasis is <-0.5; if the data does not meet the corresponding range (naked eye <5.0, spherical lens <-0.5, equivalent spherical lens after mydriasis) when submitting naked eye vision, spherical lens data and mydriasis equivalent spherical lens data, the system will pop up a prompt window to prompt that the data may be incorrect, and remind the data entry person to check and submit after it is correct. When the equivalent spherical lens after mydriasis is <-0.5, it is inferred that the student may be myopic, and the naked eye vision may be <5.0, and the spherical lens may be <-0.5.

[0052] When a student is judged to be hyperopic based on a certain indicator, other indicators may meet the corresponding conditions. For example, if the axial length is <23mm, it may be hyperopia, and the naked eye vision will not be too bad, and may be >5.0; however, it cannot be inferred that the axial length may be <23mm based on the naked eye vision >5.0, and the prompt function is the same as myopia. According to the equivalent spherical lens power after mydriasis >0.75, it can be judged that the student is hyperopic, then the student's spherical lens may be >0.75, and the naked eye vision will not be too bad, and may be >5.0. Similarly, it cannot be inferred that the equivalent spherical lens power after mydriasis >0.75 based on the naked eye vision >5.0.

[0053] S105: Correct the eye health index data to be determined that has not passed the determination to obtain the third eye health screening data.

[0054] Based on the estimated range, it is determined whether the submitted data is incorrect, and then a reminder window pops up and is sent to the on-site staff, so that the staff can manually review the eye health index data to be determined based on the reminder window, thereby eliminating the incorrect data and re-entering the data. After eliminating the incorrect data, the third eye health screening data can be obtained.

[0055] S106: Inputting the third eye health screening data into the pre-trained prediction model to screen the eye health screening data of the target to be tested.

[0056] When conducting the fourth screening, the eye health screening data of the target to be tested is screened by inputting the third eye health screening data into the pre-trained prediction model. Specifically, the pre-trained prediction model is to conduct an in-depth analysis of historical data and establish a highly accurate prediction model based on this. The core of the fourth screening is to systematically comb and mine the rich historical data accumulated in the database, and use statistical methods and machine learning techniques to construct a prediction model that can reflect the inherent laws and trends of the data. When new data is entered, these prediction models can quickly predict and evaluate the new data in real time. By comparing the predicted values ​​with the actual entered values, the system can quickly identify anomalies or deviations in the data, and give an accuracy evaluation of the entered data accordingly. This process not only helps to improve the overall quality of the database, but also provides more accurate and reliable data support for decision-making.

[0057] In one embodiment, when performing the fourth screening, part of the data can be selected from the third eye health screening data set as a prediction set, and the data other than the prediction set can be used as a comparison set; the part of the data includes basic identity information and / or eye health indicator data; the prediction set is input into a preset prediction model to obtain a prediction result; the prediction result is compared with the comparison set to obtain a comparison error; if the comparison error is greater than a preset threshold, the data corresponding to the comparison set is screened.

[0058] Furthermore, when training the prediction model, it is necessary to obtain historical eye health indicator data of eye health indicators; in addition to basic identity information and eye health indicator data, the historical eye health indicator data also include the age of myopia onset, parents' educational level, parents' myopia status, and outdoor exercise time obtained from questionnaire information; principal component analysis is performed on each indicator to obtain characteristic indicators, and the characteristic data corresponding to the characteristic indicators in the historical eye health indicator data are determined; after deleting the missing values ​​in the characteristic data, the characteristic data is divided into a training set and a test set; the pre-trained prediction model is trained using the training set, and the model performance is evaluated through cross-validation.

[0059] Specifically, the fourth screening can be divided into two parts. First, after collecting basic identity information such as age, gender, region, grade, etc., use this basic identity information to establish a data model based on statistical analysis and machine learning algorithms to predict the first eye health indicator. By comparing the model prediction value with the actual collected first indicator value, we can preliminarily evaluate the accuracy and reliability of the data and provide a basis for subsequent analysis. Secondly, after the first eye health indicator is entered, the basic identity information and the first eye health indicator that have been entered, as well as other possible eye health indicators, can be used to predict other unknown or to-be-entered indicators based on multiple regression analysis and deep learning. These predictions can be based on multiple regression analysis, deep learning or other advanced algorithms, which can comprehensively consider the interactions and influences between multiple factors. Through this step, we can further verify the integrity and consistency of the data and discover possible data anomalies or errors.

[0060] Specifically, when using basic identity information for prediction, because this basic identity information usually does not directly determine eye health-related indicators, it is necessary to use historical data information of the indicator.

[0061] (1) Dataset

[0062] The database contains historical data on gender, date of birth, ID number, school, grade, contact address, and indicators to be predicted. The original data format is as follows:

[0063] index Value gender Chinese characters, male, female Date of Birth The format is YYYY-MM-DD ID number String, 18 digits in total, consisting of 17 digits of the main body code and 1 digit of the check code School Chinese string, the full name of the student's school grade "kindergarten", "first grade", "second grade", "...", "twelfth grade" Contact Address Chinese string, representing the student's permanent address

[0064] (2) The above data sets are processed so that the indicator data meet the input requirements of the prediction model. The specific processing method is as follows:

[0065]

[0066]

[0067] Among them, when the birth date indicator is processed, when the record is entered into the system, there will be an indicator "creation time" to save the timestamp at that time. The timestamp format is YYYY-MM-DDHH:MM:SS. Subtract the year of the birth date from the year of the "creation time" indicator to get the preliminary age. Check whether the month and day of the "creation time" have passed the month and day of the birth date. If they have, there is no need to adjust the age. If they have not, then the age needs to be subtracted by 1, so as to calculate the student's birth date and age at the time of inspection based on the creation time.

[0069] According to the actual unknown of the school, determine the urban and rural category of the school location, and establish the school and urban and rural correspondence table. The format of the correspondence table is shown in the following table:

[0070]

[0071] After the school table is established, when the data entry personnel enter the name of the student's school when collecting data, the machine can automatically detect the urban and rural type of the school.

[0072] When making predictions, variables such as gender and age can be used as independent variables, and the indicator data to be predicted can be used as dependent variables. Use the decision tree regression model for regression prediction. For example, after entering the student's identity information, the quality control system can read the student's gender, age, grade, school urban and rural type, and contact address urban and rural type. The first indicator that needs to be predicted is right eye vision. Encode the existing data classification variables in the database (such as gender, grade, school urban and rural type, contact address urban and rural type), check the right eye vision data to ensure that it meets the requirements of the regression model (such as continuity), and select appropriate model parameters such as maximum depth, minimum sample split number, and minimum sample leaf node number according to data characteristics and needs. Divide the data set into a training set and a test set, usually in a ratio of 70%:30% or 80%:20%. The training set is used to train the model, and the test set is used to evaluate the model performance.

[0073] After the model training is completed, the variables such as gender, school, age, etc. of the students to be tested are substituted into the prediction model to obtain the prediction results. They are used as the reference range of the eye health indicators of the students to be tested. When the eye health indicators of the students to be tested are input, the deviation between the input results and the predicted results is calculated, and the evaluation of the results is given. When the deviation is too large, a corresponding reminder is given.

[0074] After the first eye health indicator is entered, gender, grade, urban or rural area of ​​the school is used as stratification marks (the data in the database are grouped according to gender (male and female), grade (kindergarten, grades one to six, junior high school, high school), and school (urban area, county town, and rural area). All historical data are divided into 24 groups according to the grouping labels, including urban area-kindergarten-male, county town-kindergarten-male, etc.), the entered eye health indicator is used as the independent variable, and the historical data of the indicator to be tested in the database is used as the dependent variable to establish a regression model.

[0075] Based on the historical data stored in the database, we can obtain the historical records of eye health indicators, including the age of myopia onset, as well as variables such as parents' education level, parents' myopia status, and outdoor exercise time obtained through questionnaire information. If a school lacks the corresponding questionnaire database, we will not be able to obtain specific data for these variables, so its indicator value will be considered blank.

[0076] When performing regression analysis, missing values ​​need to be processed first. Missing values ​​or outliers will have a huge impact on the prediction results, so consider deleting them directly and not incorporating them into the regression model. Divide the data set into a training set and a test set. Use the training set to train the model and adjust the model parameters to minimize the loss function. During the training process, use cross-validation and other techniques to evaluate the performance of the model and prevent overfitting. Use the test set to evaluate the performance of the model and calculate evaluation indicators such as standard error and goodness of fit. Deploy the trained model to the quality control system to predict new data. Monitor the performance of the model and update or retrain as needed. For example, after a 10-year-old boy in Primary School X has completed a vision test and a refractive status test, the boy's axial length needs to be measured. At this time, extract the vision test indicators, refractive status test indicators, and axial length data of samples with urban and rural signs of "urban areas", grade of primary school, and gender of "male" from the database as the sample set. Take the vision test indicators and refractive status test indicators as independent variables and the axial length as the result variable, and include them in the machine learning model for training. After obtaining the training model, the visual acuity test index and refractive status index of the 10-year-old boy were incorporated into the model to obtain the predicted value of the boy. After the screening staff input the actual axial measurement information, the deviation between the actual value and the regression value was calculated, and the deviation evaluation was given.

[0077] It should be noted that there is often a large correlation between eye health examination indicators. For example, when predicting the axial length of the left eye, the model includes the left eye's living visual acuity, the left eye's corrected visual acuity, the left eye's naked eye visual acuity, the left eye's spherical lens, the left eye's cylindrical lens, and the equivalent spherical lens. Including indicators with strong correlations in the model at the same time will lead to a high degree of collinearity, resulting in distorted model estimates or difficulty in accurate estimation. Therefore, before incorporating them into the model, it is necessary to perform a dimensionality reduction operation on the principal component analysis of the independent variables.

[0078] Principal component analysis reduces computational costs and improves model performance by retaining features with the largest variance. The steps for principal component analysis of independent variables are as follows:

[0079] S601: Data standardization: Since various eye health examination indicators may have different dimensions and value ranges, the original data must first be standardized to eliminate the impact of the dimension.

[0080] S602: Calculate the correlation matrix. Based on the standardized data, calculate the correlation coefficient matrix between the indicators to reflect the degree of linear correlation between them.

[0081] S603: Solve the eigenvalues ​​and eigenvectors, perform eigenvalue decomposition on the correlation coefficient matrix, and obtain a series of eigenvalues ​​and corresponding eigenvectors. These eigenvalues ​​and eigenvectors will be used to construct new principal components.

[0082] S604: Determine the number of principal components: Determine the number of principal components based on the size of the eigenvalue and the cumulative contribution rate (i.e., the ratio of the information contained in the first few principal components to the total information). The quality control model selects principal components whose cumulative contribution rate reaches a certain threshold (e.g., 85%).

[0083] S605: Constructing principal components: Using the selected eigenvectors and original data, construct new principal component variables. These principal component variables are linear combinations of the original indicators and are independent of each other.

[0084] Through principal component analysis, the original indicators with strong correlation are converted into independent principal component variables, which effectively reduces the collinearity problem in the model. And because the principal component variables contain most of the information in the original indicators and are independent of each other, the regression model constructed using principal component variables usually has higher accuracy and predictive ability.

[0085] like Figure 3 As shown, the embodiment of the present application also provides a quality control device for eye health screening data, including:

[0086] The data acquisition module 301 acquires initial eye health screening data of the target to be tested, wherein the initial eye health screening data includes basic identity information and eye health index data.

[0087] The first screening module 302 performs integrity check and accuracy check on the basic identity information to obtain first eye health screening data.

[0088] The second screening module 303 performs an indicator accuracy check on the eye health indicator data in the first eye health screening data to obtain second eye health screening data.

[0089] The indicator association module 304 obtains the value of each indicator in the second eye health screening data, and determines the eye health indicator data to be determined through the value of each indicator and the preset indicator relationship.

[0090] The third screening module 305 amends the undetermined eye health indicator data to be determined to obtain the third eye health screening data.

[0091] The prediction and screening module 306 inputs the third eye health screening data into a pre-trained prediction model to screen the eye health screening data of the target to be tested.

[0092] like Figure 4 As shown, an embodiment of the present application further provides a quality control device for eye health screening data, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can:

[0093] Obtain initial eye health screening data of the target to be measured, wherein the initial eye health screening data includes basic identity information and eye health indicator data; perform integrity check and accuracy check on the basic identity information to obtain first eye health screening data; perform indicator accuracy check on the eye health indicator data in the first eye health screening data to obtain second eye health screening data; obtain the value of each indicator in the second eye health screening data, and determine the eye health indicator data to be determined through the values ​​of each indicator and the preset indicator relationship; correct the eye health indicator data to be determined that fails the determination to obtain third eye health screening data; input the third eye health screening data into a pre-trained prediction model to screen the eye health screening data of the target to be measured.

[0094] The embodiment of the present application further provides a non-volatile computer storage medium storing computer executable instructions, wherein the computer executable instructions are configured as follows:

[0095] Obtain initial eye health screening data of the target to be measured, wherein the initial eye health screening data includes basic identity information and eye health indicator data; perform integrity check and accuracy check on the basic identity information to obtain first eye health screening data; perform indicator accuracy check on the eye health indicator data in the first eye health screening data to obtain second eye health screening data; obtain the value of each indicator in the second eye health screening data, and determine the eye health indicator data to be determined through the values ​​of each indicator and the preset indicator relationship; correct the eye health indicator data to be determined that fails the determination to obtain third eye health screening data; input the third eye health screening data into a pre-trained prediction model to screen the eye health screening data of the target to be measured.

[0096] Each embodiment in this application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device and medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0097] The devices and media provided in the embodiments of the present application correspond one-to-one to the methods. Therefore, the devices and media also have similar beneficial technical effects as the corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the devices and media will not be repeated here.

[0098] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0099] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0100] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0101] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0102] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0103] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0104] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include transitory media such as modulated data signals and carrier waves.

[0105] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0106] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A quality control method for eye health screening data, characterized in that: include: Obtaining initial eye health screening data of the target to be tested, wherein the initial eye health screening data includes basic identity information and eye health indicator data; Perform integrity check and accuracy check on the basic identity information to obtain the first eye health screening data; Performing an indicator accuracy check on the eye health indicator data in the first eye health screening data to obtain second eye health screening data; Obtaining the values ​​of each indicator in the second eye health screening data, and determining the eye health indicator data to be determined through the values ​​of each indicator and the preset indicator relationship; Correct the undetermined eye health index data to obtain the third eye health screening data; The third eye health screening data is input into the pre-trained prediction model to screen the eye health screening data of the target to be tested.

2. The method according to claim 1, characterized in that The basic identity information includes at least name, gender, date of birth, ID number, school, grade, class, contact number, and contact address; And / or, the integrity check and accuracy check of the basic identity information specifically includes: Determine the required information in the basic identity information, and determine whether the required information is complete; Use regular expressions to check the format of ID card numbers and phone numbers; Verify the contact address using the address library and check the date format of the date of birth; Determine the data source of the basic identity information, and if the data source of the basic identity information is multiple, perform consistency check on the basic identity information of the multiple data sources.

3. The method according to claim 1, characterized in that The eye health index data in the first eye health screening data include: eye axis, visual acuity, refractive status, cornea, pupil distance, intraocular pressure, accommodation amplitude, accommodation sensitivity, stereoscopic vision, latent deviation, ratio of accommodative convergence to accommodation, corneal diameter, Worth's four points, convergence near point breaking point, blur point, recovery point, positive and negative flip shooting through difficult lenses, long-distance BI examination, long-distance BO examination, near BI examination and near BO examination.

4. The method according to claim 1, characterized in that: The obtaining of the values ​​of each indicator in the second eye health screening data, and determining the eye health indicator data to be determined through the values ​​of each indicator and the preset indicator relationship, specifically includes: determining a first eye health indicator and a second eye health indicator in the second eye health screening data; The first eye health index is an eye health index that can be used to predict the eye state, and the second eye health index is an eye health index that can be predicted by the eye state; Determining a predicted value of the eye state of the target to be measured by using the first eye health indicator; Based on the eye state prediction value, determining a numerical prediction interval of a second eye health index of the target to be measured; If the indicator value corresponding to the second eye health indicator is not within the second eye health indicator value prediction interval, the indicator value corresponding to the second eye health indicator is the eye health indicator data to be determined.

5. The method according to claim 4, characterized in that The first eye health indicator includes at least one of an eye axis and an equivalent spherical diopter after mydriasis; The eye condition includes at least one of myopia and hyperopia; The second eye health indicator includes at least one of the equivalent spherical power after mydriasis, spherical lens, and naked eye vision.

6. The method according to claim 1, characterized in that The step of inputting the third eye health screening data into the pre-trained prediction model to screen the eye health screening data of the target to be tested specifically includes: Selecting part of the data in the third eye health screening data set as a prediction set, and using data other than the prediction set as a comparison set; the part of the data includes basic identity information and / or eye health index data; Inputting the prediction set into a preset prediction model to obtain a prediction result; By comparing the prediction result with the comparison set, a comparison error is obtained; If the comparison error is greater than a preset threshold, the data corresponding to the comparison set is screened.

7. The method according to claim 6, characterized in that Before inputting the prediction set into the preset prediction model, the method further includes: Obtain eye health indicator data of previous years; in addition to basic identity information and eye health indicator data, the eye health indicator data of previous years also includes the age of myopia onset, the educational level of parents obtained from questionnaire survey information, the myopia status of parents, and outdoor exercise time; Performing principal component analysis on each indicator to obtain a characteristic indicator, and determining the characteristic data corresponding to the characteristic indicator in the eye health indicator data of the previous years; After deleting missing values ​​in the feature data, the feature data is divided into a training set and a test set; The pre-trained prediction model is trained using the training set, and the model performance is evaluated by cross-validation.

8. A quality control device for eye health screening data, characterized in that: include: A data acquisition module is used to acquire initial eye health screening data of the target to be tested, wherein the initial eye health screening data includes basic identity information and eye health index data; The first screening module performs integrity check and accuracy check on the basic identity information to obtain the first eye health screening data; A second screening module performs an indicator accuracy check on the eye health indicator data in the first eye health screening data to obtain second eye health screening data; An indicator association module is used to obtain the values ​​of each indicator in the second eye health screening data, and determine the eye health indicator data to be determined through the values ​​of each indicator and the preset indicator relationship; The third screening module amends the undetermined eye health index data to be determined to obtain the third eye health screening data; The predictive screening module inputs the third eye health screening data into a pre-trained prediction model to screen the eye health screening data of the target to be tested.

9. A quality control device for eye health screening data, characterized in that: include: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute: Obtaining initial eye health screening data of the target to be tested, wherein the initial eye health screening data includes basic identity information and eye health indicator data; Perform integrity check and accuracy check on the basic identity information to obtain the first eye health screening data; Performing an indicator accuracy check on the eye health indicator data in the first eye health screening data to obtain second eye health screening data; Obtaining the values ​​of each indicator in the second eye health screening data, and determining the eye health indicator data to be determined through the values ​​of each indicator and the preset indicator relationship; Correct the undetermined eye health index data to obtain the third eye health screening data; The third eye health screening data is input into the pre-trained prediction model to screen the eye health screening data of the target to be tested.

10. A non-volatile computer storage medium storing computer executable instructions, characterized in that: The computer executable instructions are configured to: Obtaining initial eye health screening data of the target to be tested, wherein the initial eye health screening data includes basic identity information and eye health indicator data; Perform integrity check and accuracy check on the basic identity information to obtain the first eye health screening data; Performing an indicator accuracy check on the eye health indicator data in the first eye health screening data to obtain second eye health screening data; Obtaining the values ​​of each indicator in the second eye health screening data, and determining the eye health indicator data to be determined through the values ​​of each indicator and the preset indicator relationship; Correct the undetermined eye health index data to obtain the third eye health screening data; The third eye health screening data is input into the pre-trained prediction model to screen the eye health screening data of the target to be tested.

Images