Ophthalmic data analysis method, apparatus, medium based on population cohort analysis

Abstract

The application discloses an ophthalmic data analysis method, device and medium based on population cohort analysis, which comprises the following steps: constructing a first Bayesian decision network based on ophthalmic related data; calculating the Simpson index between reference nodes in each initial community, and recursively selecting each node of the first Bayesian decision network based on the Simpson index and a preset probability threshold to obtain a second Bayesian decision network; clustering the intermediate community of the second Bayesian decision network to obtain a target Bayesian decision network; and generating a target ophthalmic data analysis result based on the variable values of the nodes of each target community in the target Bayesian decision network and the conditional dependency relationship between the nodes. The application realizes automatic analysis of ophthalmic related data through population cohort analysis, and obtains a target ophthalmic data analysis result, thereby improving the accuracy of the ophthalmic data analysis result and the analysis efficiency compared with an artificial analysis scheme.

Description

Technical Field

[0001] This application relates to, but is not limited to, the field of artificial intelligence technology, and in particular to an ophthalmic data analysis method, device, and medium based on population cohort analysis. Background Technology

[0002] The causes and related factors of ophthalmic diseases are complex and diverse. To accurately understand the pathogenesis of ophthalmic diseases and develop appropriate precision treatments, it is necessary to collect detailed ophthalmic data and analyze it to obtain results that provide effective auxiliary information for the diagnosis of ophthalmic diseases. The ophthalmic data to be collected includes imaging data corresponding to ophthalmic diseases, OSDI scale data, electronic medical records, individual data from different populations (such as age, family situation, and lifestyle habits), and epidemiological data. Furthermore, according to traditional Chinese medicine theory, the eyes have rich correlations with organs throughout the body; therefore, data from other parts of the body besides the eyes, as well as treatment data, also need to be collected. This data collection and analysis process involves large medical data samples, containing measurements of hundreds of features. Manually analyzing this sample data would be extremely time-consuming, and the accuracy of the results could not be guaranteed. Summary of the Invention

[0003] This application provides an ophthalmic data analysis method, device, and medium based on population cohort analysis, which can improve the accuracy of ophthalmic data analysis results and increase analysis efficiency.

[0004] In a first aspect, embodiments of this application provide an ophthalmic data analysis method based on population cohort analysis, the method comprising:

[0005] A first Bayesian decision network is constructed based on ocular symptom types, intervention plan information, influencing factors, diagnostic indicators, patient quality of life information, and etiological information. The network form of the first Bayesian decision network is a directed acyclic graph. The first Bayesian decision network includes multiple initial communities, each initial community includes multiple nodes, and each node corresponds to a variable. The feature type of the variable is one of the ocular symptom types, intervention plan information, influencing factors, diagnostic indicators, patient quality of life information, or etiological information. The variables corresponding to two nodes connected by directed arrows have a conditional dependency relationship, and the conditional dependency relationship corresponds to a conditional probability value.

[0006] Calculate the percentage of nodes in each of the initial communities;

[0007] The Simpson index between reference nodes in each initial community is calculated based on the proportion of each node. The reference node is used to indicate two nodes in the initial community connected by a directed arrow, and the Simpson index is used to indicate the joint probability that the variables corresponding to the two reference nodes belong to the same category.

[0008] Based on the Simpson index and a preset probability threshold, each node of the first Bayesian decision network is recursively selected to obtain a second Bayesian decision network, wherein the second Bayesian decision network includes multiple intermediate communities.

[0009] Clustering is performed on all the intermediate communities to obtain the target Bayesian decision network, wherein the target Bayesian decision network includes multiple target communities;

[0010] The target ophthalmology data analysis results are generated based on the variable values ​​of the nodes of each target community in the target Bayesian decision network and the conditional dependencies between the nodes.

[0011] In some embodiments, a second Bayesian decision network is obtained by recursively selecting each node of the first Bayesian decision network based on the Simpson index and a preset probability threshold, including:

[0012] Assign a preset number of category labels to each node, and calculate the Simpson index between nodes after each label assignment.

[0013] When it is detected that the Simpson index corresponding to all reference nodes in the initial community is greater than the preset probability threshold, and the proportion of the number of nodes in each initial community is less than the preset proportion threshold, the recursive operation is stopped, and the second Bayesian decision network is obtained.

[0014] In some embodiments, clustering is performed on all the intermediate communities to obtain the target Bayesian decision network, including:

[0015] Calculate the differential effect size among all the intermediate communities;

[0016] Based on the differential effect size, the ABC analysis algorithm is used to recursively select all the intermediate communities to obtain the target Bayesian decision network.

[0017] In some embodiments, based on the differential effect size, the ABC analysis algorithm is used to recursively select all the intermediate communities to obtain the target Bayesian decision network, including:

[0018] Weights are assigned to each of the intermediate communities, and the intermediate communities are divided based on all the weights to obtain a first community set, a second community set, and a third community set. Each of the first community set, the second community set, and the third community set has at least one of the intermediate communities. The community weight corresponding to the first community set is greater than the community weight corresponding to the second community set, and the community weight corresponding to the second community set is greater than the community weight corresponding to the third community set.

[0019] The differential effect size corresponding to the intermediate community in the first type of community set, the second type of community set, and the third type of community set is determined in descending order of community weight.

[0020] When the differential effect size is detected to be less than a preset effect size threshold, a reference community corresponding to the differential effect size that is less than the preset effect size threshold is determined, and the reference communities are merged to obtain a third Bayesian decision network, wherein the reference community belongs to the intermediate community, and the number of reference communities is at least 2.

[0021] When the number of intermediate communities in the third Bayesian decision network falls within the Miller optimal range m, the third Bayesian decision network is determined as the target Bayesian decision network.

[0022] In some embodiments, the Simpson index among reference nodes in each initial community is calculated based on the proportion of each of the nodes, and is obtained according to the following formula:

[0023]

[0024] Where D is the Simpson index, S is the number of nodes in each initial community, i is the node index in the initial community, and P... i This represents the percentage of the total number of nodes.

[0025] In some embodiments, the differential effect size among all the intermediate communities is calculated according to the following formula:

[0026]

[0027] Where d is the difference effect size, M1 and M2 are the variable averages of the two intermediate communities to be compared, and SD1 is the mean value of the difference effect size. 2 SD2 2 These are the standard deviations of the variables for the two intermediate communities to be compared.

[0028] In some embodiments, after generating target ophthalmology data analysis results based on the variables of nodes in each target community in the target Bayesian decision network and the conditional dependencies between nodes, the method further includes:

[0029] Generate a visual interface;

[0030] The results of the target ophthalmological data analysis are displayed in the visualization interface.

[0031] Secondly, embodiments of this application provide a control device, including at least one control processor and a memory for communicatively connecting to the at least one control processor; the memory stores instructions executable by the at least one control processor, which, when executed by the at least one control processor, enable the at least one control processor to perform the ophthalmic data analysis method based on population cohort analysis as described in the first aspect.

[0032] Thirdly, embodiments of this application also provide a computer device, including the control device of the second aspect.

[0033] Fourthly, embodiments of this application also provide a computer-readable storage medium storing computer-executable instructions for performing the ophthalmic data analysis method based on population cohort analysis as described in the first aspect.

[0034] This application provides an ophthalmic data analysis method, device, and medium based on population cohort analysis. The method includes: constructing a first Bayesian decision network based on ocular symptom types, intervention program information, and influencing factors. The first Bayesian decision network is a directed acyclic graph (DAG). It includes multiple initial communities, each with multiple nodes. Each node corresponds to a variable, and the variable's characteristic type is one of the ocular symptom types, intervention program information, or influencing factors. Variables corresponding to two nodes connected by directed arrows have a conditional dependency relationship, and this conditional dependency relationship corresponds to a conditional probability value. The method also includes calculating the percentage of nodes in each initial community; and calculating the percentage of nodes in each initial community. The Simpson index is used to determine the joint probability that the variables corresponding to the two reference nodes belong to the same category. Based on the Simpson index and a preset probability threshold, each node in the first Bayesian decision network is recursively selected to obtain a second Bayesian decision network, which includes multiple intermediate communities. All intermediate communities are clustered to obtain a target Bayesian decision network, which includes multiple target communities. Target ophthalmology data analysis results are generated based on the variable values ​​of nodes in each target community and the conditional dependencies between nodes. According to the scheme provided in this application embodiment, automatic analysis of ophthalmology-related data is achieved through population cohort analysis to obtain target ophthalmology data analysis results. Compared with manual analysis, this improves the accuracy and efficiency of ophthalmology data analysis results. Attached Figure Description

[0035] Figure 1 This is a flowchart of the steps of an ophthalmic data analysis method based on population cohort analysis provided in one embodiment of this application;

[0036] Figure 2 This is a structural diagram of a computer device provided in another embodiment of this application. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0038] It is understandable that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, or the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0039] The causes and related factors of ophthalmic diseases are complex and diverse. To accurately understand the pathogenesis of ophthalmic diseases and develop appropriate precision treatments, it is necessary to collect detailed ophthalmic data and analyze it to obtain results that provide effective auxiliary information for the diagnosis of ophthalmic diseases. The ophthalmic data to be collected includes imaging data corresponding to ophthalmic diseases, OSDI scale data, electronic medical records, individual data from different populations (such as age, family situation, and lifestyle habits), and epidemiological data. Furthermore, according to traditional Chinese medicine theory, the eyes have rich correlations with organs throughout the body; therefore, data from other parts of the body besides the eyes, as well as treatment data, also need to be collected. This data collection and analysis process involves large medical data samples, containing measurements of hundreds of features. Manually analyzing this sample data would be extremely time-consuming, and the accuracy of the results could not be guaranteed.

[0040] To address the aforementioned problems, this application provides an ophthalmic data analysis method, device, and medium based on population cohort analysis. The method includes: constructing a first Bayesian decision network based on ocular symptom types, intervention plan information, influencing factors, diagnostic indicators, patient quality of life information, and etiological information. The first Bayesian decision network is a directed acyclic graph, comprising multiple initial communities, each initial community containing multiple nodes, and each node corresponding to a variable. The variable's characteristics are one of the ocular symptom types, intervention plan information, influencing factors, diagnostic indicators, patient quality of life information, or etiological information. Variables corresponding to two nodes connected by directed arrows have a conditional dependency relationship, and this conditional dependency relationship corresponds to a conditional probability value. The method further involves calculating the node values ​​of each initial community. The method involves calculating the Simpson index between reference nodes in each initial community based on the percentage of each node's number. A reference node indicates two nodes connected by a directed arrow in the initial community, and the Simpson index indicates the joint probability that the variables corresponding to the two reference nodes belong to the same category. Based on the Simpson index and a preset probability threshold, recursively selecting nodes in the first Bayesian decision network yields a second Bayesian decision network, which includes multiple intermediate communities. All intermediate communities are clustered to obtain a target Bayesian decision network, which includes multiple target communities. Target ophthalmology data analysis results are generated based on the variable values ​​of nodes in each target community within the target Bayesian decision network and the conditional dependencies between nodes. According to the embodiment of this application, automatic analysis of ophthalmology-related data is achieved through population cohort analysis to obtain target ophthalmology data analysis results. Compared to manual analysis, this method improves the accuracy and efficiency of ophthalmology data analysis results.

[0041] The embodiments of this application will be further described below with reference to the accompanying drawings.

[0042] refer to Figure 1 , Figure 1 This is a flowchart illustrating the steps of an ophthalmic data analysis method based on population cohort analysis, as provided in one embodiment of this application. This embodiment of the application provides an ophthalmic data analysis method based on population cohort analysis, which includes, but is not limited to, the following steps:

[0043] Step S10: Construct a first Bayesian decision network based on ocular symptom types, intervention plan information, influencing factors, diagnostic indicators, patient quality of life information, and etiological information. The first Bayesian decision network is a directed acyclic graph. The first Bayesian decision network includes multiple initial communities, each initial community includes multiple nodes, and each node corresponds to a variable. The variable's feature type is one of ocular symptom types, intervention plan information, influencing factors, diagnostic indicators, patient quality of life information, or etiological information. The variables corresponding to two nodes connected by directed arrows have a conditional dependency relationship, and the conditional dependency relationship corresponds to a conditional probability value.

[0044] Specifically, the types of eye symptoms in this embodiment include information on the types of symptoms such as stinging or burning sensation, foreign body sensation, blurred vision, eye fatigue, red eyes, itchy eyes, and abnormal tear secretion.

[0045] Specifically, the intervention information in this embodiment is intervention information for different ophthalmic symptoms, including the use of artificial tears or lubricants, warm compresses to the eyes, lacrimal duct plugging, use of cyclosporine or steroid eye drops, anti-inflammatory treatment, use of Omega-3 fatty acid supplements, acupuncture or traditional Chinese medicine treatment, and lifestyle adjustments (such as reducing screen time and increasing blinking frequency).

[0046] Specifically, the influencing factors in this embodiment are ophthalmology-related influencing factors, including age, gender, environmental factors (dry air, wind, air pollution), prolonged use of electronic devices, sleep quality, dietary habits, history of eye surgery (such as laser correction surgery), systemic diseases (such as diabetes, rheumatoid arthritis), medication use (such as antihistamines, antidepressants), and the habit of wearing contact lenses.

[0047] Specifically, the diagnostic indicators in this embodiment include tear film breakup time (BUT), corneal fluorescence staining score, tear secretion volume (Schirmer test), ocular surface inflammation markers, ocular surface temperature, ocular surface cytology examination, and ocular imaging examinations (such as fundus photography and tear film imaging).

[0048] Specifically, the patient quality of life information in this embodiment includes daily activities affected by ophthalmic diseases (such as reading and driving), visual function scores, quality of life questionnaires (OSDI, DEQ-5, etc.), and sleep quality scores.

[0049] Specifically, the causes in this embodiment include excessively rapid tear evaporation, insufficient tear secretion, meibomian gland dysfunction (MGD), and abnormal tear composition (such as abnormal lipid layer and protein content).

[0050] It is understood that the aforementioned types of ocular symptoms, intervention program information, influencing factors, diagnostic indicators, patient quality of life information, and etiological information constitute the ophthalmological association data in this embodiment. This ophthalmological association data provides an effective data foundation for constructing the first Bayesian decision network. Furthermore, the format of the first Bayesian decision network in this embodiment is a directed acyclic graph (DAG). A DAG can represent the causal relationship between multiple variables corresponding to each node in the first Bayesian decision network. Based on the first Bayesian decision network, the dependencies and structure between variables can be revealed, thereby better understanding the distribution and characteristics of the data and providing effective support for obtaining accurate target ophthalmological data analysis results in the future.

[0051] Step S20: Calculate the percentage of nodes in each initial community;

[0052] Step S30: Calculate the Simpson index between reference nodes in each initial community based on the proportion of each node. The reference node is used to indicate two nodes in the initial community connected by a directed arrow, and the Simpson index is used to indicate the joint probability that the variables corresponding to the two reference nodes belong to the same category.

[0053] Step S40: Based on the Simpson index and a preset probability threshold, recursively select each node of the first Bayesian decision network to obtain the second Bayesian decision network, wherein the second Bayesian decision network includes multiple intermediate communities.

[0054] Specifically, in this embodiment, the node count ratio is the proportion of the number of nodes in the initial community to the total number of nodes in the first Bayesian decision network.

[0055] It should be noted that in some embodiments, this embodiment calculates the Simpson index among reference nodes in each initial community based on the proportion of each node, using the following formula:

[0056]

[0057] Where D is the Simpson index, S is the number of nodes in each initial community, i is the node index in the initial community, and P... i This represents the percentage of nodes.

[0058] Understandably, since the Simpson index is used to indicate the joint probability that the variables corresponding to two reference nodes belong to the same category, the magnitude of the joint probability determines whether the two reference nodes are appropriately classified into the same community. This embodiment captures the dependencies between variables by calculating the Simpson index between reference nodes in each initial community. Based on the Simpson index and a preset probability threshold, the nodes of the first Bayesian decision network are recursively selected to obtain the second Bayesian decision network. Compared with the first Bayesian decision network, the community characteristics of the corresponding intermediate communities are more specific and the dependencies are more accurate, thus providing an effective data foundation for obtaining accurate target ophthalmological data analysis results.

[0059] Additionally, in some embodiments... Figure 1 Step S40 includes, but is not limited to, the following steps:

[0060] Step S41: Assign a preset number of category labels to each node one by one, and calculate the Simpson index between nodes after each label assignment.

[0061] Step S42: When it is detected that the Simpson index corresponding to all reference nodes in the initial community is greater than the preset probability threshold, and the proportion of the number of nodes in each initial community is less than the preset proportion threshold, the recursive operation is stopped, and the second Bayesian decision network is obtained.

[0062] Specifically, the process of obtaining the second Bayesian decision network in this embodiment is as follows: each variable of each initial community in the first Bayesian decision network is analyzed repeatedly, and the variable is placed in the recursion of the first Bayesian decision network. The variable corresponding to each node will be assigned a preset number of category labels in the recursion. The specific content of the category labels in this embodiment depends on the feature type of the corresponding node. For example, the feature type corresponding to node A is the variable of ophthalmological disease, and there are two category labels: "liver wind internal movement syndrome" and "normal". Based on the probability distribution of the current node and the probability distribution of its child nodes, the conditional probability is calculated using Bayes' theorem, and conditional dependencies are established accordingly. The relationship between different category labels of the variable is analyzed based on the Simpson index. The above process is recursively repeated. The first Bayesian decision network further subdivides the dataset according to the category label of the variable to form more specific communities until a specific condition is met: that is, when the Simpson index corresponding to all reference nodes in the initial community is greater than a preset probability threshold (i.e., all nodes in the initial community share the same category label on a specific variable), and the proportion of the number of nodes in each initial community is less than a preset proportion threshold. The preset proportion threshold in this embodiment can be selected by those skilled in the art according to the actual situation. For example, the preset proportion threshold in this embodiment is 1% of the total number of nodes in the first Bayesian decision network.

[0063] Step S50: Cluster all intermediate communities to obtain the target Bayesian decision network, wherein the target Bayesian decision network includes multiple target communities.

[0064] Specifically, in this embodiment, Figure 1 Step S50 includes, but is not limited to, the following steps:

[0065] Step S51: Calculate the differential effect size among all intermediate communities;

[0066] Step S52: Based on the differential effect size, the ABC analysis algorithm is used to recursively select all intermediate communities to obtain the target Bayesian decision network.

[0067] It is understandable that clustering all intermediate communities can reduce the number of communities generated by the Bayesian decision network. Communities with high similarity will be merged into a new community. Following the idea of ​​Fast Frugal Trees (FFTs), unnecessary decision nodes are removed to simplify the tree structure. Some accuracy is sacrificed in exchange for speed and indirectness, thereby improving the efficiency of obtaining the target ophthalmological data analysis results.

[0068] It is understood that the difference effect size in this embodiment refers to the difference caused by factors and is an indicator for measuring the magnitude of the treatment effect. By calculating the difference effect size, the contribution of each community to the target ophthalmological data analysis results can be identified and quantified. Furthermore, based on the magnitude of the difference effect size, the most relevant community is recursively selected. In other words, the difference effect size in this embodiment is used to evaluate the correlation between the intermediate community and the target ophthalmological data analysis results. This ensures that the subsequent determination of the target ophthalmological data analysis results can focus on the most meaningful data and improve the accuracy of the target ophthalmological data analysis results.

[0069] In some embodiments, the differential effect size among all intermediate communities is calculated according to the following formula:

[0070]

[0071] Where d is the difference effect size, M1 and M2 are the variable means of the two intermediate communities to be compared, and SD1 is the difference effect size. 2 SD2 2 These are the standard deviations of the variables for the two intermediate communities to be compared.

[0072] Additionally, in some embodiments... Figure 1 Step S52 includes, but is not limited to, the following steps:

[0073] Step S521: Assign weights to each intermediate community and divide the intermediate communities based on all weights to obtain a first community set, a second community set, and a third community set. Each of the first community set, the second community set, and the third community set has at least one intermediate community. The community weight corresponding to the first community set is greater than the community weight corresponding to the second community set, and the community weight corresponding to the second community set is greater than the community weight corresponding to the third community set.

[0074] Step S522: Determine the differential effect size corresponding to the intermediate community in the first type of community set, the second type of community set, and the third type of community set in descending order of community weight.

[0075] Step S523: When the differential effect size is detected to be less than the preset effect size threshold, the reference community corresponding to the differential effect size less than the preset effect size threshold is determined, the reference communities are merged to obtain the third Bayesian decision network, wherein the reference community belongs to the intermediate community, and the number of reference communities is at least 2.

[0076] Step S524: When the number of intermediate communities in the third Bayesian decision network belongs to the Miller optimal m range, the third Bayesian decision network is determined as the target Bayesian decision network.

[0077] Understandably, this embodiment uses the ABC analysis algorithm to recursively select all intermediate communities. Specifically, according to the community weights from largest to smallest, the differential effect size corresponding to each intermediate community in the first, second, and third community sets is determined sequentially. When the differential effect size is detected to be less than a preset effect size threshold, the reference community corresponding to the differential effect size less than the preset effect size threshold is determined. The reference community corresponding to the differential effect size less than the preset effect size threshold indicates that the difference between the two communities is small, and there is no need to divide them into two categories. The reference communities are merged to obtain the third Bayesian decision network. When the number of intermediate communities in the third Bayesian decision network is detected to belong to the Miller optimal m range (i.e., the number of intermediate communities is 7±2), the third Bayesian decision network is determined as the target Bayesian decision network. The target Bayesian decision network includes the most relevant target community.

[0078] Specifically, in this embodiment, the difference effect size d is divided into three levels according to its value: small (value less than 0.2), medium (fluctuating around 0.5), and large (value greater than 0.8). The larger the difference effect size, the greater the correlation (influence) of the intermediate community on the target ophthalmology data analysis results. At the same time, the difference effect size reflects the degree of difference in the influence of different variables on the two intermediate communities.

[0079] Specifically, the ABC analysis algorithm of this embodiment is a classification method used to divide all intermediate communities into three categories according to their importance: (1) Category A: the most important intermediate communities, accounting for 80% of the total value and 20% of the total number; (2) Category B: moderately important intermediate communities, accounting for 15% of the total value and 30% of the total number; Category C: the least important intermediate communities, accounting for 5% of the total value and 50% of the total number. By recursively selecting important intermediate communities using ABC analysis, it is ensured that important communities are within a manageable range (this embodiment uses Miller's optimal m range to determine this). This can effectively allocate resources, prioritize the processing of Category A communities, improve resource utilization efficiency, and aim to screen out a few communities with sufficiently obvious characteristics and greater analytical value, thereby ensuring the accuracy of the target ophthalmological data analysis results.

[0080] Step S60: Generate target ophthalmology data analysis results based on the variable values ​​of nodes in each target community in the target Bayesian decision network and the conditional dependencies between nodes.

[0081] Understandably, given the complex and diverse causes and related factors of ophthalmic diseases, this embodiment utilizes interpretable artificial intelligence technology to conduct population cohort analysis on high-dimensional medical big data containing hundreds of features in order to accurately identify key related influencing factors, obtain target ophthalmic data analysis results, and thus provide effective auxiliary information for the diagnosis of ophthalmic diseases.

[0082] Additionally, in some embodiments, during execution Figure 1 Following step S60, the ophthalmic data analysis method based on population cohort analysis provided in this application embodiment further includes, but is not limited to, the following steps:

[0083] Step S71: Generate a visual interface;

[0084] Step S72: Display the target ophthalmology data analysis results in the visualization interface.

[0085] It is understandable that after obtaining the variable values ​​of nodes based on each target community and the conditional dependencies between nodes to generate the target ophthalmology data analysis results, this embodiment uses the VisPanel visualization tool to display the target ophthalmology data analysis results in the visualization interface. This enables experts or doctors to intuitively understand and analyze the data, better grasp the data characteristics, and provide effective auxiliary information for the diagnosis of ophthalmological diseases.

[0086] like Figure 2 As shown, Figure 2 This is a structural diagram of a control device provided in one embodiment of this application. The present invention also provides a control device 200, comprising:

[0087] The processor 210 can be implemented using a general-purpose central processing unit (CPU), microprocessor, application specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application.

[0088] The memory 220 can be implemented as a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM). The memory 220 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 220 and is called and executed by the processor 210 to execute the ophthalmic data analysis method based on population cohort analysis of the embodiments of this application.

[0089] Input / output interface 230 is used to implement information input and output;

[0090] The communication interface 240 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0091] Bus 250 transmits information between various components of the device (e.g., processor 210, memory 220, input / output interface 230, and communication interface 240);

[0092] The processor 210, memory 220, input / output interface 230 and communication interface 240 are connected to each other within the device via bus 250.

[0093] In addition, this application also provides a computer device, including the control device 200 of the above embodiments.

[0094] In addition, this application embodiment also provides a storage medium, which is a computer-readable storage medium, storing a computer program that, when executed by a processor, implements the above-described ophthalmic data analysis method based on population cohort analysis.

[0095] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof. The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separate, and may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0096] It will be understood by those skilled in the art that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically include computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0097] The above provides a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of the present invention.

Claims

1. A method for ophthalmic data analysis based on population cohort analysis, characterized in that, include: A first Bayesian decision network is constructed based on ocular symptom types, intervention plan information, influencing factors, diagnostic indicators, patient quality of life information, and etiological information. The network form of the first Bayesian decision network is a directed acyclic graph. The first Bayesian decision network includes multiple initial communities, each initial community includes multiple nodes, and each node corresponds to a variable. The feature type of the variable is one of the ocular symptom types, intervention plan information, influencing factors, diagnostic indicators, patient quality of life information, or etiological information. The variables corresponding to two nodes connected by directed arrows have a conditional dependency relationship, and the conditional dependency relationship corresponds to a conditional probability value. Calculate the percentage of nodes in each of the initial communities; The Simpson index between reference nodes in each initial community is calculated based on the proportion of each node. The reference node is used to indicate two nodes in the initial community connected by a directed arrow, and the Simpson index is used to indicate the joint probability that the variables corresponding to the two reference nodes belong to the same category. Based on the Simpson index and a preset probability threshold, each node of the first Bayesian decision network is recursively selected to obtain a second Bayesian decision network, wherein the second Bayesian decision network includes multiple intermediate communities. Clustering is performed on all the intermediate communities to obtain the target Bayesian decision network, wherein the target Bayesian decision network includes multiple target communities; The target ophthalmology data analysis results are generated based on the variable values ​​of the nodes of each target community in the target Bayesian decision network and the conditional dependencies between the nodes.

2. The ophthalmic data analysis method based on population cohort analysis according to claim 1, characterized in that, Based on the Simpson index and a preset probability threshold, recursively selects each node of the first Bayesian decision network to obtain a second Bayesian decision network, including: Assign a preset number of category labels to each node, and calculate the Simpson index between nodes after each label assignment. When it is detected that the Simpson index corresponding to all reference nodes in the initial community is greater than the preset probability threshold, and the proportion of the number of nodes in each initial community is less than the preset proportion threshold, the recursive operation is stopped, and the second Bayesian decision network is obtained.

3. The ophthalmic data analysis method based on population cohort analysis according to claim 1, characterized in that, Clustering is performed on all the intermediate communities to obtain the target Bayesian decision network, including: Calculate the differential effect size among all the intermediate communities; Based on the differential effect size, the ABC analysis algorithm is used to recursively select all the intermediate communities to obtain the target Bayesian decision network.

4. The ophthalmic data analysis method based on population cohort analysis according to claim 3, characterized in that, Based on the aforementioned differential effect size, the ABC analysis algorithm is used to recursively select from all the intermediate communities to obtain the target Bayesian decision network, including: Weights are assigned to each of the intermediate communities, and the intermediate communities are divided based on all the weights to obtain a first community set, a second community set, and a third community set. Each of the first community set, the second community set, and the third community set has at least one of the intermediate communities. The community weight corresponding to the first community set is greater than the community weight corresponding to the second community set, and the community weight corresponding to the second community set is greater than the community weight corresponding to the third community set. The differential effect size corresponding to the intermediate community in the first type of community set, the second type of community set, and the third type of community set is determined in descending order of community weight. When the differential effect size is detected to be less than a preset effect size threshold, a reference community corresponding to the differential effect size that is less than the preset effect size threshold is determined, and the reference communities are merged to obtain a third Bayesian decision network, wherein the reference community belongs to the intermediate community, and the number of reference communities is at least 2. When the number of intermediate communities in the third Bayesian decision network falls within the Miller optimal range m, the third Bayesian decision network is determined as the target Bayesian decision network.

5. The ophthalmic data analysis method based on population cohort analysis according to claim 1, characterized in that, The Simpson index among reference nodes in each initial community is calculated based on the proportion of each node, using the following formula: Where D is the Simpson index, S is the number of nodes in each initial community, i is the node index in the initial community, and P... i This represents the percentage of the total number of nodes.

6. The ophthalmic data analysis method based on population cohort analysis according to claim 3, characterized in that, The differential effect size among all the intermediate communities is calculated using the following formula: Where d is the difference effect size, M1 and M2 are the variable averages of the two intermediate communities to be compared, and SD1 is the mean value of the difference effect size. 2 SD2 2 These are the standard deviations of the variables for the two intermediate communities to be compared.

7. The ophthalmic data analysis method based on population cohort analysis according to claim 1, characterized in that, After generating the target ophthalmology data analysis results based on the variables of each node in the target community in the target Bayesian decision network and the conditional dependencies between nodes, the method further includes: Generate a visual interface; The results of the target ophthalmological data analysis are displayed in the visualization interface.

8. A control device, characterized in that, It includes at least one control processor and a memory for communicatively connecting to the at least one control processor; the memory stores instructions executable by the at least one control processor, which, when executed by the at least one control processor, enable the at least one control processor to perform the ophthalmic data analysis method based on population cohort analysis as described in any one of claims 1 to 7.

9. A computer device, characterized in that, Includes the control device as described in claim 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions for causing a computer to perform the ophthalmic data analysis method based on population cohort analysis as described in any one of claims 1 to 7.

Images