Vision risk prediction method, system, device and medium based on multidimensional data
By constructing an optimized fuzzy hierarchy model and vision risk state model, combining vision status, physiological and non-physiological data, the problem of unconsidered multi-factorial combined effects in vision risk prediction in the existing technology is solved, and the accuracy and quantification of vision risk is achieved. Analyze, improves the accuracy of predictions.
Patent Information
- Application Number
- CN202311834918.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-12-27
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2043-12-27
AI Technical Summary
The existing vision risk prediction methods mainly rely on single-dimensional data and cannot fully consider multiple influencing factors, resulting in prediction errors and lack of universality.
By collecting vision status data, physiological data and non-physiological data, an optimized fuzzy hierarchy model and vision risk status model are constructed, qualitative and quantitative analysis are carried out, and risk coefficients and vision risk prediction models are generated, considering the combined role of multiple influencing factors.
Accurate and quantitative analysis of vision risks is achieved, the prediction error and lack of universality in the prior art are overcome, and more accurate vision risks are provided.
Smart Images

Figure CN117789984B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of vision risk prediction, and in particular to a method, system, device and medium for vision risk prediction based on multidimensional data. Background Art
[0002] In recent years, the incidence of myopia in my country has shown a clear upward trend, and myopia has become a major public health issue affecting the eye health of Chinese people, especially young people. This most common eye disease is listed as one of the three major diseases in the world. The incidence of myopia in my country is about 33%, which is 1.5 times the world average.
[0003] Existing vision risk prediction schemes mainly make predictions based on a single dimension, such as the physiological data of the eye. However, methods based on single-dimensional data for vision risk prediction usually have a lag, and can only see the simple correlation between single-dimensional data and vision changes, but cannot see the comprehensive influencing factors that lead to vision changes.
[0004] The prior art (invention patent with application number 2022112280578) discloses a myopia prevention and control system and method based on multi-source data, which obtains a myopia risk coefficient by counting eye behavior, determines whether myopia prediction is needed by the myopia risk coefficient, obtains a myopia prediction model through reasonable training, performs periodic myopia prediction, and provides users with prevention and control plans in a timely manner to avoid the deterioration of myopia. However, the prior art simply calculates the risk coefficient by counting eye behavior and its corresponding eye time, and the number of times the eye time threshold is exceeded. First, eye behavior and eye time have different tolerance thresholds for different groups of people. It is not universal to use a unified eye time threshold for vision risk prediction, which will cause errors in the prediction; secondly, the influencing factors are multi-source and not limited to eye use. The prior art does not consider the impact of multiple influencing factors on vision; furthermore, the influencing factors do not have a single impact on vision, but have a common impact.
[0005] In summary, the prior art lacks a vision risk prediction method that qualitatively and quantitatively considers multiple influencing factors in vision risk prediction. Summary of the invention
[0006] Based on the technical defects raised by the above-mentioned background technology, the purpose of the present invention is to provide a vision risk prediction method, system, device and medium based on multidimensional data, and to construct a data set for training, testing and optimizing fuzzy hierarchical models and vision risk status models based on changes in vision status data, physiological data and non-physiological data, thereby completing qualitative and quantitative analysis of changes in vision status using non-physiological data and physiological data, and solving the problem in the prior art of lack of a vision risk prediction method that qualitatively and quantitatively considers multiple influencing factors in vision risk prediction.
[0007] The present invention is achieved through the following technical solutions:
[0008] The first aspect of the present invention provides a method for predicting vision risk based on multidimensional data, comprising the following steps:
[0009] S1, collecting vision status data, physiological data and non-physiological data, performing change analysis on the vision status data, the physiological data and the non-physiological data, and dividing the vision status data, physiological data and non-physiological data after the change analysis into a data set including a training set and a test set;
[0010] S2. Constructing an optimized fuzzy hierarchical model and a vision risk state model, and using the training set to train the optimized fuzzy hierarchical model and the vision risk state model to generate a vision risk prediction model; wherein the training process is as follows:
[0011] Inputting the training set into the optimized fuzzy hierarchical model for training to obtain a risk coefficient;
[0012] Inputting the risk coefficient and the training set into the vision risk state model for training to generate a vision risk prediction model;
[0013] S3, using the test set to test the vision risk prediction model, if the test passes, executing step S4, if the test fails, re-executing steps S1 and S2 according to the test results;
[0014] S4. Obtain the data to be predicted, use the vision risk prediction model that has passed the test to predict the data to obtain a vision risk index, and determine the myopia risk according to the vision risk index.
[0015] In the above technical scheme, by comprehensively analyzing the vision status data, physiological data and non-physiological data, a data set for training and testing the optimized fuzzy hierarchical model and the vision risk status model is constructed based on the data changes of the vision status data, physiological data and non-physiological data. At the same time, an optimized fuzzy hierarchical model and a vision risk status model are constructed. Through the optimized fuzzy hierarchical model, qualitative and quantitative analysis of the changes in vision status of non-physiological data and physiological data can be realized, so as to obtain the risk proportion of physiological data and non-physiological data in the process of vision change, that is, the risk coefficient in this scheme. Based on the risk coefficients of physiological data and non-physiological data, the vision change process is simulated by using the vision risk status model, and finally the vision risk situation under the influence of various factors can be determined, so as to construct a vision risk prediction model for vision risk prediction. In vision risk prediction, a vision risk prediction method that qualitatively and quantitatively considers multiple influencing factors is provided, which overcomes the defects of the prior art proposed in the background technology.
[0016] In an optional embodiment, performing change analysis on the vision status data, the physiological data and the non-physiological data comprises the following steps:
[0017] Performing time synchronization processing on the vision status data, the physiological data and the non-physiological data to generate time synchronization data;
[0018] Dividing the time synchronization data using the vision status data as a standard time period to generate a plurality of vision data segments;
[0019] The changes in each vision data segment are calculated separately, and the changes are integrated to obtain the vision status change data.
[0020] In an optional embodiment, constructing an optimized fuzzy hierarchical model includes the following steps:
[0021] The visual state is set as the first level, the physiological index is set as the second level, and the non-physiological index is set as the third level;
[0022] Constructing an influence function, determining the relative importance of indicators at each level through the influence function, and constructing a judgment matrix based on the relative importance using a consistent matrix method, wherein the judgment matrix is used to determine the risk weights of physiological indicators and non-physiological indicators;
[0023] The vision status is set as a level set, the risk weights of physiological indicators and non-physiological indicators are set as a weight set, and the physiological indicators and non-physiological indicators are set as a factor set;
[0024] Constructing a membership function according to the result of the change analysis, and determining the membership degree of the factor set according to the membership function;
[0025] A risk coefficient function is constructed based on the membership degree and the risk weight.
[0026] In an optional embodiment, constructing an influence function, and determining the relative importance of indicators at each level through the influence function includes:
[0027] Constructing a first influence function, determining a physiological influence index of a physiological indicator on a visual state through the first influence function, and calculating the relative importance of each physiological indicator according to the physiological influence index;
[0028] A second influence function is constructed, a non-physiological influence index of the non-physiological indicator on the physiological indicator is determined by the second influence function, and the relative importance of each non-physiological indicator is calculated according to the non-physiological influence index.
[0029] In an optional embodiment, the first influence function is constructed as follows:
[0030]
[0031] In the above formula, represents the ith physiological index pf i For the physiological impact index of vision status v, t represents the number of segments of vision data, C j (*) represents the rate of change of data in the jth visual acuity data segment.
[0032] In an optional embodiment, the membership function is constructed as follows:
[0033]
[0034] In the above formula, U i Indicates the membership degree of index i, In t De represents the growth of index i in the tth period, t It represents the decrease of the indicator in the tth period, V represents the change of vision status, T1 represents the total number of increase indicators, and T2 represents the total number of decrease indicators.
[0035] In an optional embodiment, constructing a vision risk status model includes:
[0036] Calculate the same frequency of physiological influence between various physiological data under the same vision state, and the same frequency of non-physiological influence between various non-physiological data under the same physiological data;
[0037] A frequency influence function of the physiological influence at the same frequency and the non-physiological influence at the same frequency is constructed, and a frequency influence factor is obtained through the frequency influence function.
[0038] The second aspect of the present invention provides a vision risk prediction system based on multidimensional data, comprising:
[0039] A data processing module, the data processing module is used to collect vision status data, physiological data and non-physiological data, perform change analysis on the vision status data, the physiological data and the non-physiological data, and divide the vision status data, the physiological data and the non-physiological data after the change analysis into a data set including a training set and a test set;
[0040] A model building module, the model building module is used to build an optimized fuzzy hierarchical model and a vision risk state model, and use the training set to train the optimized fuzzy hierarchical model and the vision risk state model to generate a vision risk prediction model; wherein the training process is as follows:
[0041] Inputting the training set into the optimized fuzzy hierarchical model for training to obtain a risk coefficient;
[0042] Inputting the risk coefficient and the training set into the vision risk state model for training to generate a vision risk prediction model;
[0043] A model testing module, the model testing module is used to test the vision risk prediction model using the test set;
[0044] The risk prediction module is used to obtain the data to be predicted, use the vision risk prediction model that has passed the test to predict the data to be predicted, obtain the vision risk index, and determine the myopia risk according to the vision risk index.
[0045] A third aspect of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements a method for predicting vision risk based on multidimensional data when executing the program.
[0046] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for predicting vision risk based on multidimensional data.
[0047] Compared with the prior art, the present invention has the following advantages and beneficial effects:
[0048] 1. Based on the changes in vision status data, physiological data and non-physiological data, a data set for training, testing and optimizing fuzzy hierarchical models and vision risk status models is constructed, which overcomes the problem that the existing technology uses subjective analysis, is not universal, and will cause prediction errors;
[0049] 2. We built an optimized fuzzy hierarchical model and a vision risk status model, trained and tested them, completed the qualitative and quantitative analysis of non-physiological data and physiological data on changes in vision status, and determined the quantitative data of various influencing factors on vision risk;
[0050] 3. The frequency of influence between various factors is taken into consideration to construct a frequency influence function, thereby obtaining the influence of multiple factors on a single factor, which makes up for the fact that the existing technology does not consider the influence of multiple factors at the same time. BRIEF DESCRIPTION OF THE DRAWINGS
[0051] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:
[0052] Figure 1 A schematic diagram of a flow chart of a method for predicting vision risk based on multidimensional data provided in Example 1 of the present invention;
[0053] Figure 2 A schematic diagram of the structure of a vision risk prediction system based on multidimensional data provided in Example 2 of the present invention;
[0054] Figure 3 A schematic diagram of the structure of an electronic device provided in Example 3 of the present invention. DETAILED DESCRIPTION
[0055] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.
[0056] Example 1
[0057] Figure 1 A schematic diagram of a process for predicting vision risk based on multidimensional data provided in Example 1 of the present invention, such as Figure 1 As shown, the vision risk prediction method based on multidimensional data includes the following steps:
[0058] S1, collecting vision status data, physiological data and non-physiological data, performing change analysis on the vision status data, the physiological data and the non-physiological data, and dividing the vision status data, physiological data and non-physiological data after the change analysis into a data set including a training set and a test set;
[0059] S2. Constructing an optimized fuzzy hierarchical model and a vision risk state model, and using the training set to train the optimized fuzzy hierarchical model and the vision risk state model to generate a vision risk prediction model; wherein the training process is as follows:
[0060] Inputting the training set into the optimized fuzzy hierarchical model for training to obtain a risk coefficient;
[0061] Inputting the risk coefficient and the training set into the vision risk state model for training to generate a vision risk prediction model;
[0062] S3, using the test set to test the vision risk prediction model, if the test passes, executing step S4, if the test fails, re-executing steps S1 and S2 according to the test results;
[0063] S4. Obtain the data to be predicted, use the vision risk prediction model that has passed the test to predict the data to obtain a vision risk index, and determine the myopia risk according to the vision risk index.
[0064] It should be noted that the vision risk prediction method based on multidimensional data provided in the embodiment of the present invention is to comprehensively analyze the vision status data, physiological data and non-physiological data, and to construct a data set for training and testing the optimized fuzzy hierarchical model and the vision risk status model based on the data changes of the vision status data, physiological data and non-physiological data. At the same time, an optimized fuzzy hierarchical model and a vision risk status model are constructed. The optimized fuzzy hierarchical model can realize the qualitative and quantitative analysis of the changes in vision status of non-physiological data and physiological data, so as to obtain the risk proportion occupied by physiological data and non-physiological data in the process of vision change, that is, the risk coefficient in this scheme. Based on the risk coefficients of physiological data and non-physiological data, the vision change process is simulated by using the vision risk status model, and finally the vision risk situation under the influence of various factors can be determined, so as to construct a vision risk prediction model for vision risk prediction, and provide a vision risk prediction method that qualitatively and quantitatively considers multiple influencing factors in vision risk prediction, which overcomes the defects of the prior art proposed in the background technology.
[0065] It should be emphasized that the prior art simply calculates the risk factor by counting the eye behaviors and their corresponding eye usage time, and the number of times the eye usage time threshold is exceeded. In the prior art, the threshold is usually determined by the experience of experts, which is usually subjective. In order to avoid subjectivity, the present invention conducts qualitative and quantitative analysis of the correlation between the data based on the changes between the data. Based on this concept, the present invention performs a change analysis on the collected vision status data, physiological data, and non-physiological data, and trains and tests the subsequently constructed model based on the data after the change analysis.
[0066] In an optional embodiment, performing change analysis on the vision status data, the physiological data and the non-physiological data comprises the following steps:
[0067] Establishing a time axis, placing the vision status data, the physiological data and the non-physiological data into the time axis in the form of corresponding time, and generating time synchronization data;
[0068] Dividing the time synchronization data using the vision status data as a standard time period to generate a plurality of vision data segments;
[0069] The changes in each vision data segment are calculated separately, and the changes are integrated to obtain the vision status change data.
[0070] It should be noted that the vision status data is the feedback of the vision status, which is a result, and the core point of the present invention is to study the influence of various physiological data and non-physiological data between two vision states on the result. Therefore, in the present invention, the vision status data is used as the standard time period for division to generate several vision data segments. The changes of each vision data segment are calculated separately, and the changes of each vision data segment are analyzed, which is not only conducive to determining the influence of each influencing factor on the vision status; at the same time, it is also conducive to the analysis of the coupling relationship between high-dimensional data.
[0071] Specifically, time synchronization may be: establishing a time axis, placing vision status data, physiological data and non-physiological data into the time axis in the form of corresponding time, and generating time synchronization data.
[0072] In an optional embodiment, constructing an optimized fuzzy hierarchical model includes the following steps:
[0073] The visual state is set as the first level, the physiological index is set as the second level, and the non-physiological index is set as the third level;
[0074] Constructing an influence function, determining the relative importance of indicators at each level through the influence function, and constructing a judgment matrix based on the relative importance using a consistent matrix method, wherein the judgment matrix is used to determine the risk weights of physiological indicators and non-physiological indicators;
[0075] The vision status is set as a level set, the risk weights of the physiological status indicators and the non-physiological status indicators are set as a weight set, and the physiological indicators and the non-physiological indicators are set as a factor set;
[0076] Constructing a membership function according to the result of the change analysis, and determining the membership degree of the factor set according to the membership function;
[0077] A risk coefficient function is constructed based on the membership degree and the risk weight.
[0078] It should be emphasized that the key to qualitatively and quantitatively analyzing the correlation between the data based on the changes between the data is to find the relationship between the data. In vision testing, the vision is determined by testing the physiological indicators, and the non-physiological indicators indirectly affect the vision by affecting the physiological indicators. For example, long-term use of the eyes among non-physiological factors will lead to changes in intraocular pressure among physiological factors, and long-term changes in intraocular pressure will cause a decrease in vision. Therefore, the core point of the present invention is to find the correlation between vision and physiological indicators, and between physiological indicators and non-physiological indicators.
[0079] Specifically, the optimized fuzzy hierarchical model includes three levels, which are the third level, the second level and the first level from the bottom to the top. In the embodiment of the present invention, the visual state is set as the first level, the physiological index is set as the second level, and the non-physiological index is set as the third level. The purpose of such design is to find the correlation relationship between the levels.
[0080] In the present invention, to find the correlation between each level, it is first necessary to determine the relative importance of each indicator in each level. Usually, the relative importance of each indicator is determined by ophthalmologists based on experience. For example, experts set the importance of eye environment, eye habits, living habits, and genetic factors to 0.15, 0.4, 0.4, and 0.05, respectively. The relative importance of eye habits and living habits is 1. However, although this method is obtained by ophthalmologists based on the summary of various empirical data, the relative importance obtained is different due to the different cases and medical environments contacted by the experts. Therefore, in the present invention, the form of constructing an influence function is adopted to determine the influence index between the indicators at each level by the influence function.
[0081] The implementation process in this embodiment includes:
[0082] Constructing a first influence function, determining a physiological influence index of a physiological indicator on a visual state through the first influence function, and calculating the relative importance of each physiological indicator according to the physiological influence index;
[0083] A second influence function is constructed, a non-physiological influence index of the non-physiological indicator on the physiological indicator is determined by the second influence function, and the relative importance of each non-physiological indicator is calculated according to the non-physiological influence index.
[0084] Specifically, the first influence function is constructed as follows:
[0085]
[0086] In the above formula, represents the ith physiological index pf i For the physiological impact index of vision status v, t represents the number of segments of vision data, C j (*) represents the rate of change of data in the jth visual acuity data segment.
[0087] It should be noted that, in this embodiment, the rate of change of the data of the physiological indicator in the vision data segment and the rate of change of the data of the vision status in the vision data segment are slopes.
[0088] The second influence function is constructed as follows:
[0089]
[0090] In the above formula, Represents the kth non-physiological index npf k For the i-th physiological index pf i Non-physiological impact index.
[0091] It should be noted that, in the process of constructing the influence function, the present invention focuses on the relationship between the change rate of each indicator at a lower level and the affected indicator at the upper level, rather than the relationship between quantities.
[0092] After obtaining each physiological impact index and each non-physiological impact index, the relative importance needs to be calculated. In this embodiment, the physiological impact index and the non-physiological impact index are calculated using the advantage analysis method to obtain the relative importance.
[0093] Specifically, assuming that there are n physiological indicators in total, k physiological indicators are randomly selected from the n physiological indicators to generate a set, and the proportion of the k physiological indicators in the set is calculated respectively, and the proportion is used as the contribution value of the corresponding physiological indicator. Taking physiological indicators as an example, the formula for calculating relative importance using the advantage analysis method is as follows:
[0094]
[0095] In the above formula, relpf i Indicates the relative importance of the ith physiological index, (Inf pfi,v ) k Represents the contribution value of the i-th physiological index.
[0096] The judgment matrix of physiological factors is an n*n two-dimensional matrix, where the matrix elements in the i-th row and j-th column can be expressed by the formula Similarly, the matrix element in the jth row and kth column can also be calculated using the above formula, and the matrix needs to satisfy a ij *a jk =a ik , thereby realizing the construction of the judgment matrix, completing the importance judgment of each lower-level indicator under the corresponding upper-level indicator, and initially obtaining an initial risk weight.
[0097] The calculation method of the relative importance of non-physiological indicators and the construction method of the judgment matrix are similar to those of physiological indicators. The relative importance of non-physiological indicators and the judgment matrix can be derived through the above examples.
[0098] In an optional embodiment, the membership function is constructed as follows:
[0099]
[0100] In the above formula, U i Indicates the membership degree of index i, In t De represents the growth of index i in the tth period, t It represents the decrease of the indicator in the tth period, V represents the change of vision status, T1 represents the total number of increase indicators, and T2 represents the total number of decrease indicators.
[0101] It should be noted that the present invention constructs a membership function according to the change of each data, which is specifically manifested as, when the degree of vision risk increases, there is an increase in a certain index, and the increase in the index is greater than the decrease in the index. Therefore, in the present invention, two sets are constructed for the increasing index and the decreasing index, namely, the increasing index set and the decreasing index set, respectively, and the increasing amount and the decreasing amount in the set are comprehensively calculated respectively, so as to obtain the membership of the index to the change of vision status.
[0102] In this embodiment, the risk coefficient function is a function that multiplies the membership degree by the risk weight.
[0103] In an optional embodiment, constructing a vision risk status model includes:
[0104] Calculate the same frequency of physiological influence between various physiological data under the same vision state, and the same frequency of non-physiological influence between various non-physiological data under the same physiological data;
[0105] A frequency influence function of the physiological influence at the same frequency and the non-physiological influence at the same frequency is constructed, and a frequency influence factor is obtained through the frequency influence function.
[0106] It should be emphasized that the influencing factors do not affect vision alone, but have a common influence. The prior art has not considered the many-to-one nature of the influencing factors. For example, the effect of long-term use of salt in a dim environment on vision is greater than long-term use of eyes in a bright environment. Therefore, the present invention considers the influence of the same frequency between the factors, and constructs a frequency influence function to obtain the influence of multiple factors on a single factor.
[0107] In this embodiment, the same frequency of physiological influence between various physiological data under the same vision state is realized as follows: under the same vision state, a data set is formed for two or more physiological data among the optional physiological data, and the ratio between the data in the data set is calculated. If the ratio is equal to or approximately the ratio of the risk coefficient obtained by the optimized fuzzy hierarchical model, it is considered to be same-frequency data, otherwise it is non- same-frequency data. The ratio of the same-frequency data to the total data is calculated as the same frequency of physiological influence of multiple physiological data in the data set.
[0108] Example 2
[0109] Figure 2 This is a schematic diagram of the structure of a vision risk prediction system based on multidimensional data provided in Example 2 of the present invention, such as Figure 3 As shown, the vision risk prediction system based on multidimensional data includes:
[0110] A data processing module, the data processing module is used to collect vision status data, physiological data and non-physiological data, perform change analysis on the vision status data, the physiological data and the non-physiological data, and divide the vision status data, the physiological data and the non-physiological data after the change analysis into a data set including a training set and a test set;
[0111] A model building module, the model building module is used to build an optimized fuzzy hierarchical model and a vision risk state model, and use the training set to train the optimized fuzzy hierarchical model and the vision risk state model to generate a vision risk prediction model; wherein the training process is as follows:
[0112] Inputting the training set into the optimized fuzzy hierarchical model for training to obtain a risk coefficient;
[0113] Inputting the risk coefficient and the training set into the vision risk state model for training to generate a vision risk prediction model;
[0114] A model testing module, the model testing module is used to test the vision risk prediction model using the test set;
[0115] The risk prediction module is used to obtain the data to be predicted, use the vision risk prediction model that has passed the test to predict the data to be predicted, obtain the vision risk index, and determine the myopia risk according to the vision risk index.
[0116] Example 3
[0117] Figure 3 A schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention is shown in FIG. Figure 3 As shown, the electronic device includes a processor 21, a memory 22, an input device 23 and an output device 24; the number of processors 21 in the computer device can be one or more. Figure 3 A processor 21 is taken as an example; the processor 21, the memory 22, the input device 23 and the output device 24 in the electronic device can be connected by a bus or other means. Figure 3 The example of connecting through bus is taken in the following.
[0118] The memory 22 is a computer-readable storage medium that can be used to store software programs, computer executable programs, and modules. The processor 21 executes various functional applications and data processing of the electronic device by running the software programs, instructions, and modules stored in the memory 22, that is, the vision risk prediction method based on multidimensional data of Example 1 is implemented.
[0119] The memory 22 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory 22 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 22 may further include a memory remotely arranged relative to the processor 21, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.
[0120] The input device 23 can be used to receive the ID and password input by the user, etc. The output device 24 is used to output the network configuration page.
[0121] Example 4
[0122] Embodiment 4 of the present invention further provides a computer-readable storage medium, wherein the computer-executable instructions, when executed by a computer processor, are used to implement the vision risk prediction method based on multidimensional data as provided in Embodiment 1.
[0123] An embodiment of the present invention provides a storage medium containing computer executable instructions, and its computer executable instructions are not limited to the method operations provided in Example 1, and can also execute related operations in the vision risk prediction method based on multidimensional data provided in any embodiment of the present invention.
[0124] The specific implementation methods 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 implementation method 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 vision risk prediction method based on multidimensional data, characterized in that: The steps include: S1, collecting vision status data, physiological data and non-physiological data, performing change analysis on the vision status data, the physiological data and the non-physiological data, and dividing the vision status data, physiological data and non-physiological data after the change analysis into a data set including a training set and a test set; S2, constructing an optimized fuzzy hierarchical model and a vision risk state model, using the training set to train the optimized fuzzy hierarchical model and the vision risk state model to generate a vision risk prediction model; wherein the training process is as follows: Inputting the training set into the optimized fuzzy hierarchical model for training to obtain a risk coefficient; Inputting the risk coefficient and the training set into the vision risk state model for training to generate a vision risk prediction model; S3, using the test set to test the vision risk prediction model, if the test passes, executing step S4, if the test fails, re-dividing the data and executing steps S1 and S2; S4, obtaining data to be predicted, using the vision risk prediction model that has passed the test to predict the data to be predicted, obtaining a vision risk index, and determining the myopia risk according to the vision risk index; Performing change analysis on the vision status data, the physiological data and the non-physiological data comprises the following steps: Performing time synchronization processing on the vision status data, the physiological data and the non-physiological data to generate time synchronization data; Dividing the time synchronization data using the vision status data as a standard time period to generate a plurality of vision data segments; Calculate the change of each vision data segment respectively, and integrate the change to obtain vision status change data; The construction of optimized fuzzy hierarchical model includes the following steps: The visual state is set as the first level, the physiological index is set as the second level, and the non-physiological index is set as the third level; Constructing an influence function, determining the relative importance of indicators at each level through the influence function, and constructing a judgment matrix based on the relative importance using a consistent matrix method, wherein the judgment matrix is used to determine the risk weights of physiological indicators and non-physiological indicators; The vision status is set as a level set, the risk weights of physiological indicators and non-physiological indicators are set as a weight set, and the physiological indicators and non-physiological indicators are set as a factor set; Constructing a membership function according to the result of the change analysis, and determining the membership degree of the factor set according to the membership function; Constructing a risk coefficient function based on the membership degree and the risk weight; Constructing an influence function, through which the relative importance of indicators at each level is determined includes: Constructing a first influence function, determining a physiological influence index of a physiological indicator on a visual state through the first influence function, and calculating the relative importance of each physiological indicator according to the physiological influence index; Constructing a second influence function, determining a non-physiological influence index of a non-physiological indicator on a physiological indicator through the second influence function, and calculating the relative importance of each non-physiological indicator according to the non-physiological influence index; The first influence function is constructed as follows: In the above formula, Represents the i-th physiological index For visual status The physiological impact index, t represents the number of visual acuity data segments, Represents the rate of change of data in the jth segment of vision data.
2. The method for predicting vision risk based on multidimensional data according to claim 1, characterized in that: The membership function is constructed as follows: In the above formula, represents the degree of membership of index i, It represents the growth of index i in the tth period, represents the decrease of the indicator in the tth period, V represents the change in vision status, represents the total number of growth indicators, Indicates the total number of reduction indicators.
3. The method for predicting vision risk based on multidimensional data according to claim 1, characterized in that: Modeling vision risk status includes: Calculate the same frequency of physiological influence between various physiological data under the same vision state, and the same frequency of non-physiological influence between various non-physiological data under the same physiological data; A frequency influence function of the physiological influence at the same frequency and the non-physiological influence at the same frequency is constructed, and a frequency influence factor is obtained through the frequency influence function.
4. A vision risk prediction system for implementing the vision risk prediction method based on multidimensional data as claimed in any one of claims 1 to 3, characterized in that: include: A data processing module, the data processing module is used to collect vision status data, physiological data and non-physiological data, perform change analysis on the vision status data, the physiological data and the non-physiological data, and divide the vision status data, the physiological data and the non-physiological data after the change analysis into a data set including a training set and a test set; A model building module, the model building module is used to build an optimized fuzzy hierarchical model and a vision risk state model, and use the training set to train the optimized fuzzy hierarchical model and the vision risk state model to generate a vision risk prediction model; wherein the training process is as follows: Inputting the training set into the optimized fuzzy hierarchical model for training to obtain a risk coefficient; Inputting the risk coefficient and the training set into the vision risk state model for training to generate a vision risk prediction model; A model testing module, the model testing module is used to test the vision risk prediction model using the test set; The risk prediction module is used to obtain the data to be predicted, use the vision risk prediction model that has passed the test to predict the data to be predicted, obtain the vision risk index, and determine the myopia risk according to the vision risk index.
5. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for predicting vision risk based on multidimensional data as claimed in any one of claims 1 to 3 is implemented.
6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for predicting vision risk based on multidimensional data as described in any one of claims 1 to 3 is implemented.
Citation Information
Patent Citations
Method, device and system for predicting myopia risk in children
CN107358036A
Myopia risk assessment method, device and system based on artificial intelligence
CN112365107A
Deep learning-based teenager myopia prediction method and model
CN117059269A
Identifying and ranking risk factors using trained predictive models
US20170323075A1
Systems and methods for hybrid record unification using a combination of deterministic, probabilistic, and possibilistic operations
US20230244697A1
