Intelligent infectious disease education content generation system

Through the intelligent infectious disease education content generation system based on the BERT model, combining pathogen identification and user interaction data, instant update and targeted adjustment of content are achieved, solving the problem of untimely update of content in traditional systems and improving the timeliness of content and learning effects.

CN118942723AInactive Publication Date: 2024-11-12SHANGHAI LINGLI HEALTH MANAGEMENT CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410782984.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-18
Publication Date
2024-11-12
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The traditional intelligent infectious disease education content generation system lacks immediate updates and targeted adjustments, resulting in the content being unable to timely reflect the latest research and data in the context of rapidly changing epidemic situation, affecting the relevance and effectiveness of educational content, and failing to be effective. Optimize learning paths using users' historical interactive data.

Method used

Using an intelligent infectious disease education content generation system based on the BERT model, the model is initialized through the medical pre-training module, combining pathogen identification and transmission pathway analysis, knowledge integration module to integrate epidemiological research results and statistical data, and the content custom module analyzes user interaction data. Adjust the difficulty and type of content, and update the model parameters in real time through the online learning and feedback module to optimize content generation.

Benefits of technology

It improves the professionalism and accuracy of content generation, enables educational materials to quickly reflect changes in the epidemic, enhances the timeliness and practicality of the content, meets the needs of different learners, and greatly improves the learning effect.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118942723A_ABST
    Figure CN118942723A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of content generation, in particular to an intelligent infectious disease education content generation system, which comprises a medical pre-training module, a knowledge integration module, a content customization module and an online learning and feedback module, and is characterized in that the medical pre-training module initializes a BERT model based on a medical literature and an infectious disease data set; the method comprises the following steps: adjusting medical terms and epidemiological concepts, and inputting key data of pathogen recognition and transmission path analysis to obtain a training knowledge structure chart; according to the method, medical terms and epidemiological concepts are accurately analyzed by adjusting the natural language processing model, the professional property and accuracy of content generation are improved, latest epidemiological research achievements and prevention and treatment information are dynamically integrated into the content, so that educational materials can quickly reflect epidemic situation changes, and the epidemic prevention and treatment efficiency is improved. The timeliness and practicability of the content are enhanced, the education content is adjusted in a targeted mode by analyzing the user interaction history, the requirements of different learners are met, and the learning effect is greatly improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of content generation, and in particular to an intelligent infectious disease education content generation system. Background Art

[0002] The field of content generation technology involves the automatic or semi-automatic creation of text, images, audio and video content, and is widely used in industries such as media, advertising, education and entertainment. In the field of education, content generation technology is particularly important, providing customized teaching materials to help learners receive education in an adaptive way. This technology can analyze educational needs and generate content that meets specific educational goals and learner needs. In addition, content generation technology can also achieve rapid conversion of multi-language and multi-format content, improving the accessibility and inclusiveness of learning.

[0003] Among them, the intelligent infectious disease education content generation system is an educational tool that aims to enhance the public's awareness and understanding of infectious diseases. By analyzing current infectious disease data and trends, it automatically generates relevant educational content such as articles, illustrations and videos. Its uses include school education, public health promotion and personal health management. Especially during global or regional infectious disease outbreaks, the system can quickly provide the latest and scientific epidemic prevention knowledge and guidance to help people correctly respond to infectious disease threats, thereby improving public health and safety.

[0004] Traditional systems lack instant updates and targeted adjustments for specific situations, resulting in the inability of content to reflect the latest research and data in a timely manner in the rapidly changing epidemic context, thus affecting the relevance and effectiveness of educational content. Traditional systems fail to effectively utilize users' historical interaction data to optimize learning paths, resulting in content that is inconsistent with users' actual needs, limiting the maximization of educational effects. For example, during an infectious disease outbreak, the lack of instantly updated educational materials and targeted learning support will delay the understanding and response to important health information and reduce the actual impact of educational measures. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an intelligent infectious disease education content generation system.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: an intelligent infectious disease education content generation system, the system comprising:

[0007] The medical pre-training module initializes the BERT model based on medical literature and infectious disease datasets, adjusts medical terminology and epidemiological concepts, inputs key data on pathogen identification and transmission pathway analysis, and obtains a training knowledge structure diagram;

[0008] The knowledge integration module collects epidemiological research results and statistical data based on the training knowledge structure diagram, integrates prevention and control information and epidemic dynamic information in a structured format, and adjusts the data hierarchy to generate an epidemic knowledge system;

[0009] The content customization module analyzes the user's historical interaction data, including the correct answer rate and reading preference, based on the epidemic knowledge system, and adjusts the content according to the analysis results, pathogen identification and prevention information, including adjusting the content difficulty and education type, and generating educational content configuration;

[0010] The online learning and feedback module is based on the educational content configuration and user feedback data, including interaction and learning effects, to perform real-time updates of model parameters, and combines the current epidemic dynamics to iteratively optimize the content generation process and generate content optimization results.

[0011] The present invention is improved in that the initialization steps of the BERT model are specifically as follows:

[0012] Based on medical literature and infectious disease datasets, the BERT model is pre-trained using text datasets;

[0013] The BERT model parameters are adjusted for medical terms and epidemiological concepts using the formula:

[0014]

[0015] Get the adjusted model parameters W′, where W is the original model parameter, α is the learning rate, δL / δW is the gradient of the loss L with respect to the parameter W, and g i is the gradient of the ith sample, λ is the gradient adjustment factor, which optimizes the sensitivity of the model to the target data, σ is the standard deviation of the sample gradient, which is used to normalize the gradient contribution, and n is the number of samples in the batch.

[0016] The present invention is improved in that the steps of inputting the pathogen identification and transmission pathway analysis data are specifically as follows:

[0017] Collect pathogen raw data from public health databases and hospital logs, including infection cases, symptom descriptions, and geographic location information;

[0018] Screening data related to pathogen identification and transmission pathways, including pathogen type, infection frequency, and transmission pathways, to obtain targeted data sets;

[0019] Input the targeted dataset into the adjusted BERT model through the formula:

[0020]

[0021] Evaluate the information dispersion S of the targeted data set and obtain the training knowledge structure diagram, where x i is the eigenvalue of the i-th data point, representing the pathogen information, μ is the average eigenvalue of the data, used to measure the central tendency, N is the total number of data points, indicating the number of samples, and p is the weighted exponent, used to amplify the impact of errors.

[0022] The present invention is improved in that the steps of collecting the epidemiological research results and statistical data are specifically as follows:

[0023] Based on the training knowledge structure diagram, identify key epidemiological research and public health databases, and use data extraction technology to extract data points related to the epidemic, including infection rate, transmission speed, and intervention effect;

[0024] Using the data weighting formula:

[0025]

[0026] Get the weighted data set D new , and integrate them to generate epidemiological research results and statistical data sets, among which D i is the value of the i-th data point, representing epidemic statistics or research results, W i is the data point weight, which is set according to the timeliness, accuracy and relevance of the data. C is the constant adjustment term used to balance the baseline impact of the data point. β is the power optimization term used to adjust the influence of the data weight. Z is the adjustment parameter of the normalization factor, and α is the power of the normalization adjustment.

[0027] The present invention is improved in that the steps of acquiring the epidemic knowledge system are specifically as follows:

[0028] Filter prevention and control information and epidemic dynamics from the epidemiological research results and statistical data, and integrate and classify the filtered information;

[0029] The preliminary integrated information is weighted using the formula:

[0030]

[0031] Get weighted information list I weighted , where I i is the original value of the i-th information, W i is the weight determined according to the urgency and criticality of the information, K is the baseline adjustment constant, γ is the weighted emphasis factor used to optimize the impact of the information weight, L is the adjustment constant, T is the total number of information items, and δ is the adjustment factor used to adjust the normalized influence;

[0032] Optimize the data hierarchy based on the weighted information list and generate the epidemic knowledge system.

[0033] The present invention is improved in that the analysis step of the user historical interaction data is specifically as follows:

[0034] Based on the epidemic knowledge system, collect historical interaction data of users on the education platform, including answer accuracy and reading preferences;

[0035] Perform statistical analysis on user interaction data and calculate the average accuracy and preference distribution of users using the formula:

[0036]

[0037] Get the weighted average accuracy and preference score R new , where r i represents the accuracy rate of the user’s i-th interaction, p i represents the corresponding preference weight, C i is the baseline correction factor for each interaction used to adjust the preference weights, γ is the weighted emphasis factor, δ is the normalized adjustment factor used to balance the preference weights, ξ is the total adjustment index, and p j is the weight of the jth interaction instance.

[0038] The present invention is improved in that the step of obtaining the educational content configuration is specifically as follows:

[0039] Based on the analysis results of the user's historical interaction data, evaluate the user's analysis depth and interest in the differentiated educational content to obtain user demand analysis data;

[0040] Analyze data based on user needs, adjust the difficulty and type of educational content, and use the formula:

[0041]

[0042] Get the adjusted educational content configuration C new , where C old For the current educational content configuration, including difficulty and education type, D target is the target difficulty, D actual is the user's mastery level, based on the user behavior analysis results, T diff is the difference between the target and education type, λ is the adjustment factor, m is the constant used to adjust the baseline, ∈ is the adjustment coefficient used to control the dependence of the adjustment result on the number of samples, and n is the number of user interactions, which is used for normalization.

[0043] The present invention is improved in that the steps of obtaining the content optimization results are specifically as follows:

[0044] Based on the educational content configuration, combined with user feedback data and current epidemic dynamics, the parameters of the educational content generation model are updated in real time to obtain updated model parameters;

[0045] Apply the updated model parameters, combine user interaction and learning effects, and iteratively optimize the content using the formula:

[0046]

[0047] Generate educational content optimization results, among which O old is the previous educational content configuration, ΔO represents the content change caused by the new parameters, λ is the optimization rate, ω is the weight of the content change, ρ is the baseline adjustment amount for a smooth update process, μ is a normalization constant used to maintain the stability of the content after update, and σ is an adjustment factor used to adjust the sensitivity of content changes.

[0048] Compared with the prior art, the advantages and positive effects of the present invention are:

[0049] In the present invention, by adjusting the natural language processing model to accurately parse medical terms and epidemiological concepts, the professionalism and accuracy of content generation are improved, and the latest epidemiological research results and prevention and control information are dynamically integrated into the content, so that educational materials can quickly reflect changes in the epidemic situation, enhancing the timeliness and practicality of the content. By analyzing the user interaction history to adjust the educational content in a targeted manner, the needs of different learners can be met, greatly improving the learning effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A module diagram of an intelligent infectious disease education content generation system is proposed for the present invention;

[0051] Figure 2 The present invention proposes an initialization flow chart of the BERT model in an intelligent infectious disease education content generation system;

[0052] Figure 3 The present invention proposes a flow chart for inputting pathogen identification and transmission pathway analysis data in an intelligent infectious disease education content generation system;

[0053] Figure 4 A flow chart for collecting epidemiological research results and statistical data in an intelligent infectious disease education content generation system is proposed for the present invention;

[0054] Figure 5 The present invention proposes a flow chart for obtaining the epidemic knowledge system in an intelligent infectious disease education content generation system;

[0055] Figure 6 The present invention proposes a flow chart for analyzing historical user interaction data in an intelligent infectious disease education content generation system;

[0056] Figure 7The present invention proposes a flow chart for obtaining educational content configuration in an intelligent infectious disease educational content generation system;

[0057] Figure 8 The present invention proposes a flow chart for obtaining content optimization results in an intelligent infectious disease education content generation system. DETAILED DESCRIPTION

[0058] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0059] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating positions or positional relationships, are based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0060] Example

[0061] See also Figure 1 The present invention provides a technical solution: an intelligent infectious disease education content generation system comprising:

[0062] The medical pre-training module initializes the BERT model based on medical literature and infectious disease datasets, adjusts medical terminology and epidemiological concepts, inputs key data on pathogen identification and transmission pathway analysis, and obtains a training knowledge structure diagram;

[0063] The knowledge integration module collects epidemiological research results and statistical data based on the training knowledge structure diagram, integrates prevention and control information and epidemic dynamics information in a structured format, and adjusts the data hierarchy to generate an epidemic knowledge system;

[0064] The content customization module is based on the epidemic knowledge system and analyzes the user's historical interaction data, including the correct answer rate and reading preferences. It adjusts the content according to the analysis results, pathogen identification and prevention information, including adjusting the content difficulty and education type, and generating educational content configuration;

[0065] The online learning and feedback module is based on educational content configuration and updates model parameters in real time according to user feedback data, including interaction and learning effects. It also iteratively optimizes the content generation process in combination with current epidemic dynamics to generate content optimization results.

[0066] The training knowledge structure diagram includes term index information, concept standardization, and data links. The epidemic knowledge system includes epidemic prevention plans, real-time statistical information, and education frameworks. The education content configuration includes difficulty levels, content types, and interactive feedback analysis results. Content optimization results include parameter adjustment effects, satisfaction indicators, and effect evaluation information.

[0067] See also Figure 2 , the initialization steps of the BERT model are as follows:

[0068] Based on medical literature and infectious disease datasets, the BERT model is pre-trained using text datasets;

[0069] The BERT model parameters are adjusted for medical terms and epidemiological concepts using the formula:

[0070]

[0071] Get the adjusted model parameters W′, where W is the original model parameter, α is the learning rate, δL / δW is the gradient of the loss L with respect to the parameter W, and g i is the gradient of the ith sample, λ is the gradient adjustment factor, which optimizes the sensitivity of the model to the target data, σ is the standard deviation of the sample gradient, which is used to normalize the gradient contribution, and n is the number of samples in the batch.

[0072] Assume the following data:

[0073] Suppose there are 3 samples, and the gradient g of each sample is 0.5, 1.5, and 0.3 respectively;

[0074] σ is calculated as the standard deviation of the three sample gradients;

[0075] Gradient of the loss function Assumed to be 0.02;

[0076] α is set to 0.01;

[0077] λ is set to 0.1;

[0078] W is the original model parameter, assumed to be a weight matrix or weight vector.

[0079] Calculation process:

[0080] First calculate σ:

[0081]

[0082] σ≈0.525

[0083] Next, calculate the weighted gradient term corresponding to each sample:

[0084]

[0085] Accumulate the above weighted gradients:

[0086] Total gradient = weighted gradient 1 + Weighted Gradient 2 + Weighted Gradient 3 =0.0219+0.0257+

[0087] 0.0211=0.0687

[0088] Finally, update W according to the learning rate:

[0089] W′=W+0.01×0.0687=W+0.000687

[0090] See also Figure 3 ,The input steps of pathogen identification and transmission pathway analysis data are as follows:

[0091] Collect pathogen raw data from public health databases and hospital logs, including infection cases, symptom descriptions, and geographic location information;

[0092] Screening data related to pathogen identification and transmission pathways, including pathogen type, infection frequency, and transmission pathways, to obtain targeted data sets;

[0093] Input the targeted dataset into the adjusted BERT model through the formula:

[0094]

[0095] Evaluate the information dispersion S of the targeted data set and obtain the training knowledge structure diagram, where x i is the eigenvalue of the i-th data point, representing the pathogen information, μ is the average eigenvalue of the data, used to measure the central tendency, N is the total number of data points, indicating the number of samples, and p is the weighted exponent, used to amplify the impact of errors.

[0096] Assume the following values:

[0097] Assume that the data point x i Indicates infection rate: x = [10%, 20%, 15%, 25%, 30%], converted to decimal form: x = [0.1, 0.2, 0.15, 0.25, 0.3];

[0098] Assume p=2 (for calculating variance), N=5 (total number of data points).

[0099] Calculate the mean μ:

[0100]

[0101] Compute the square of the weighted deviation for each data point:

[0102] (x 1 -μ) 2 =(0.1-0.2) 2 =0.01

[0103] (x 2 -μ) 2 =(0.2-0.2) 2 =0.00

[0104] (x 3 -μ) 2 =(0.15-0.2) 2 =0.0025

[0105] (x 4 -μ) 2 =(0.25-0.2) 2 =0.0025

[0106] (x 5 -μ) 2 =(0.3-0.2) 2 =0.01

[0107] Compute the weighted average of the squared deviations:

[0108]

[0109] Calculate the final S value:

[0110]

[0111] See also Figure 4 , the specific steps for collecting epidemiological research results and statistical data are:

[0112] Based on the training knowledge structure map, identify key epidemiological research and public health databases, and use data extraction technology to extract data points related to the epidemic, including infection rate, transmission speed, and intervention effect;

[0113] Using the data weighting formula:

[0114]

[0115] Get the weighted data set D new , and integrate them to generate epidemiological research results and statistical data sets, among which D i is the value of the i-th data point, representing epidemic statistics or research results, W iis the data point weight, which is set according to the timeliness, accuracy and relevance of the data. C is the constant adjustment term used to balance the baseline impact of the data point. β is the power optimization term used to adjust the influence of the data weight. Z is the adjustment parameter of the normalization factor, and α is the power of the normalization adjustment.

[0116] Assume the following data:

[0117] Data point D = [0.1, 0.2, 0.15] (e.g., representing infection rate);

[0118] Weight W = [1, 2, 1.5];

[0119] Constant C = 0.05;

[0120] Power β = 2;

[0121] Normalization factor Z = 4.5 (sum of weights);

[0122] Normalization adjustment α=1.5.

[0123] Compute the weighted sum for each data point:

[0124]

[0125] Apply a normalization factor:

[0126]

[0127] See also Figure 5 , the specific steps for acquiring the epidemic knowledge system are:

[0128] Filter prevention and control information and epidemic dynamics from epidemiological research results and statistical data, and integrate and classify the filtered information;

[0129] The preliminary integrated information is weighted using the formula:

[0130]

[0131] Get weighted information list I weighted , where I i is the original value of the i-th information, W i is the weight determined according to the urgency and criticality of the information, K is the baseline adjustment constant, γ is the weighted emphasis factor used to optimize the impact of the information weight, L is the adjustment constant, T is the total number of information items, and δ is the adjustment factor used to adjust the normalized influence;

[0132] Optimize the data hierarchy based on the weighted information list and generate the epidemic knowledge system.

[0133] Assume the following values:

[0134] Assume there are 3 data points I = [0.1, 0.2, 0.15];

[0135] Weight W = [1, 2, 1.5];

[0136] Baseline enhancement constant K = 0.05;

[0137] Power γ = 2;

[0138] Adjustment constant L = 0.1;

[0139] The total number of information items T = 3;

[0140] Normalization adjustment δ=0.5.

[0141] Denominator calculation:

[0142] L+T δ =0.1+3 0.5 ≈0.1+1.732=1.832

[0143] Calculation of each weighted term:

[0144] For I 1 =0.1,W 1 =1:

[0145]

[0146] For I 2 =0.2,W 2 =2:

[0147]

[0148] For I 3 =0.15,W 3 =1.5:

[0149]

[0150] Sum:

[0151] Total weighted sum = 0.0123 + 0.1105 + 0.0413 ≈ 0.1641

[0152] See also Figure 6 ,The specific steps for analyzing user historical interaction data are:

[0153] Based on the epidemic knowledge system, collect users' historical interaction data on the education platform, including answer accuracy and reading preferences;

[0154] Perform statistical analysis on user interaction data and calculate the average accuracy and preference distribution of users using the formula:

[0155]

[0156] Get the weighted average accuracy and preference score R new , where r i represents the accuracy rate of the user’s i-th interaction, p i represents the corresponding preference weight, C i is the baseline correction factor for each interaction used to adjust the preference weights, γ is the weighted emphasis factor, δ is the normalized adjustment factor used to balance the preference weights, ξ is the total adjustment index, and p j is the weight of the jth interaction instance.

[0157] Assume the following values:

[0158] Number of interactions n = 3;

[0159] Correct rate r = [0.8, 0.5, 0.9];

[0160] Preference weight p = [2, 1, 3];

[0161] Baseline correction C = [0.05, 0.05, 0.05];

[0162] Weighted emphasis γ = 2;

[0163] Normalization adjustment δ = 2;

[0164] The total adjustment ξ=1.5.

[0165] Compute the weighted sum:

[0166]

[0167] Calculate the normalized denominator:

[0168]

[0169] Apply total adjustments:

[0170]

[0171] Calculate the weighted average accuracy and preference score R using the given parameters and data points new Approximately 4.034.

[0172] See also Figure 7 , the specific steps for obtaining educational content configuration are:

[0173] Based on the analysis results of user historical interaction data, evaluate the user's analysis depth and interest in differentiated educational content, and obtain user demand analysis data;

[0174] Analyze data based on user needs, adjust the difficulty and type of educational content, and use the formula:

[0175]

[0176] Get the adjusted educational content configuration C new , where C old For the current educational content configuration, including difficulty and education type, D target is the target difficulty, D actual is the user's mastery level, based on the user behavior analysis results, T diff is the difference between the target and education type, λ is the adjustment factor, m is the constant used to adjust the baseline, ∈ is the adjustment coefficient used to control the dependence of the adjustment result on the number of samples, and n is the number of user interactions, which is used for normalization.

[0177] Assume the following values:

[0178] Number of user interactions n = 3;

[0179] Current Configuration C old =5;

[0180] Target Difficulty D target =7;

[0181] Practical Mastery D actual =4;

[0182] Type Difference T diff =2;

[0183] Adjustment factor λ = 0.1;

[0184] Constant m = 1;

[0185] The adjustment coefficient ∈=0.5.

[0186] Compute the weighted sum for each data point:

[0187] Molecular part = (C old ·(D target -D actual )·T diff +λ)=5·(7-4)·2+0.1=31.1

[0188]

[0189] Calculate the normalized denominator:

[0190] Denominator=1+∈·n=1+0.5·3=2.5

[0191] Calculate the final result:

[0192]

[0193] See also Figure 8 , the specific steps to obtain content optimization results are:

[0194] Based on the configuration of educational content, combined with user feedback data and the current epidemic situation, the parameters of the educational content generation model are updated in real time to obtain updated model parameters;

[0195] Apply the updated model parameters, combine user interaction and learning effects, and iteratively optimize the content using the formula:

[0196]

[0197] Generate educational content optimization results, among which O old is the previous educational content configuration, ΔO represents the content change caused by the new parameters, λ is the optimization rate, ω is the weight of the content change, ρ is the baseline adjustment amount for a smooth update process, μ is a normalization constant used to maintain the stability of the content after update, and σ is an adjustment factor used to adjust the sensitivity of content changes.

[0198] Assume the following values:

[0199] O old =50, initial content configuration score;

[0200] ΔO = 5, significant changes implemented based on user feedback;

[0201] λ = 0.1, using a moderate optimization rate to avoid mutation;

[0202] ω = 2, high impact of content change;

[0203] ρ = 1, small baseline adjustment;

[0204] μ = 10, stable denominator;

[0205] σ=0.5, balances the sensitivity of content changes.

[0206] Calculate the numerator:

[0207] Molecule = O old +λ·(ω·ΔO+ρ)=50+0.1·(2·5+1)=50+0.1·(10+1)

[0208] =50+1.1=51.1

[0209] Calculate the denominator:

[0210]

[0211] Calculate the final result:

[0212]

[0213] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. An intelligent infectious disease education content generation system, It is characterized in that The system comprises: The medical pre-training module initializes the BERT model based on medical literature and infectious disease datasets, adjusts medical terminology and epidemiological concepts, inputs key data on pathogen identification and transmission pathway analysis, and obtains a training knowledge structure diagram; The knowledge integration module collects epidemiological research results and statistical data based on the training knowledge structure diagram, integrates prevention and control information and epidemic dynamic information in a structured format, and adjusts the data hierarchy to generate an epidemic knowledge system; The content customization module analyzes the user's historical interaction data, including the correct answer rate and reading preference, based on the epidemic knowledge system, and adjusts the content according to the analysis results, pathogen identification and prevention information, including adjusting the content difficulty and education type, and generating educational content configuration; The online learning and feedback module is based on the educational content configuration and user feedback data, including interaction and learning effects, to perform real-time updates of model parameters, and combines the current epidemic dynamics to iteratively optimize the content generation process and generate content optimization results.

2. The intelligent infectious disease education content generation system according to claim 1, It is characterized in that The initialization steps of the BERT model are specifically as follows: Based on medical literature and infectious disease datasets, the BERT model is pre-trained using text datasets; The BERT model parameters are adjusted for medical terms and epidemiological concepts using the formula: Get the adjusted model parameters W′, where W is the original model parameter, a is the learning rate, δL / δW is the gradient of the loss L with respect to the parameter W, and g i is the gradient of the ith sample, λ is the gradient adjustment factor, which optimizes the sensitivity of the model to the target data, σ is the standard deviation of the sample gradient, which is used to normalize the gradient contribution, and n is the number of samples in the batch.

3. The intelligent infectious disease education content generation system according to claim 1, It is characterized in that The steps for inputting the pathogen identification and transmission pathway analysis data are specifically as follows: Collect pathogen raw data from public health databases and hospital logs, including infection cases, symptom descriptions, and geographic location information; Screening data related to pathogen identification and transmission pathways, including pathogen type, infection frequency, and transmission pathways, to obtain targeted data sets; Input the targeted dataset into the adjusted BERT model through the formula: Evaluate the information dispersion S of the targeted data set and obtain the training knowledge structure diagram, where x i is the eigenvalue of the i-th data point, representing the pathogen information, μ is the average eigenvalue of the data, used to measure the central tendency, N is the total number of data points, indicating the number of samples, and p is the weighted exponent, used to amplify the impact of errors.

4. The intelligent infectious disease education content generation system according to claim 1, It is characterized in that The specific steps for collecting the epidemiological research results and statistical data are as follows: Based on the training knowledge structure diagram, identify key epidemiological research and public health databases, and use data extraction technology to extract data points related to the epidemic, including infection rate, transmission speed, and intervention effect; Using the data weighting formula: Get the weighted data set D new , and integrate them to generate epidemiological research results and statistical data sets, among which D i is the value of the i-th data point, representing epidemic statistics or research results, W i is the data point weight, which is set according to the timeliness, accuracy and relevance of the data. C is the constant adjustment term used to balance the baseline impact of the data point. β is the power optimization term used to adjust the influence of the data weight. Z is the adjustment parameter of the normalization factor, and α is the power of the normalization adjustment.

5. The intelligent infectious disease education content generation system according to claim 1, It is characterized in that The specific steps for obtaining the epidemic knowledge system are as follows: Filter prevention and control information and epidemic dynamics from the epidemiological research results and statistical data, and integrate and classify the filtered information; The preliminary integrated information is weighted using the formula: Get weighted information list I weighted , where I i is the original value of the i-th information, W i is the weight determined according to the urgency and criticality of the information, K is the baseline adjustment constant, γ is the weighted emphasis factor used to optimize the impact of the information weight, L is the adjustment constant, T is the total number of information items, and δ is the adjustment factor used to adjust the normalized influence; Optimize the data hierarchy based on the weighted information list and generate the epidemic knowledge system.

6. The intelligent infectious disease education content generation system according to claim 1, It is characterized in that The analysis steps of the user historical interaction data are specifically as follows: Based on the epidemic knowledge system, collect historical interaction data of users on the education platform, including answer accuracy and reading preferences; Perform statistical analysis on user interaction data and calculate the average accuracy and preference distribution of users using the formula: Get the weighted average accuracy and preference score R new , where r i represents the accuracy rate of the user’s i-th interaction, p i represents the corresponding preference weight, C i is the baseline correction factor for each interaction used to adjust the preference weights, γ is the weighted emphasis factor, δ is the normalized adjustment factor used to balance the preference weights, ξ is the total adjustment index, and p j is the weight of the jth interaction instance.

7. The intelligent infectious disease education content generation system according to claim 1, It is characterized in that The steps for obtaining the educational content configuration are specifically as follows: Based on the analysis results of the user's historical interaction data, evaluate the user's analysis depth and interest in the differentiated educational content to obtain user demand analysis data; Analyze data based on user needs, adjust the difficulty and type of educational content, and use the formula: Get the adjusted educational content configuration C new , where C old For the current educational content configuration, including difficulty and education type, D target is the target difficulty, D actual is the user's mastery level, based on the user behavior analysis results, T diff is the difference between the target and education type, λ is the adjustment factor, m is the constant used to adjust the baseline, ∈ is the adjustment coefficient used to control the dependence of the adjustment result on the number of samples, and n is the number of user interactions, which is used for normalization.

8. The intelligent infectious disease education content generation system according to claim 1, It is characterized in that The steps for obtaining the content optimization results are specifically as follows: Based on the educational content configuration, combined with user feedback data and current epidemic dynamics, the parameters of the educational content generation model are updated in real time to obtain updated model parameters; Apply the updated model parameters, combine user interaction and learning effects, and iteratively optimize the content using the formula: Generate educational content optimization results, among which O old is the previous educational content configuration, ΔO represents the content change caused by the new parameters, λ is the optimization rate, ω is the weight of the content change, ρ is the baseline adjustment amount for a smooth update process, μ is a normalization constant used to maintain the stability of the content after update, and σ is an adjustment factor used to adjust the sensitivity of content changes.