A system and method for collecting and importing disease database data based on dynamic forms

Abstract

The present invention discloses a system and method for collecting and importing data of a special disease database based on a dynamic form, comprising a data collection and organization module, a subject database data set creation module, a dynamic form design and management module, a special disease database data configuration management module, a special disease database data retrieval module, and a data visualization module. The data collection and organization module is used to collect user input data, the subject database data set creation module is used to create different subject data sets, the dynamic form design and management module is used for form design and visualization management, the special disease database data configuration management module is used for configuring data and mining analysis, the special disease database data retrieval module is used for data retrieval, and the data display and analysis module is used for visualization display. The present invention discloses a system and method for collecting and importing data of a special disease database based on a dynamic form, proposes a dynamic form design algorithm based on a probability model to design a dynamic form, and proposes a special disease database data mining algorithm based on a neural network to perform data mining analysis.

Description

Technical Field

[0001] The present invention relates to the fields of form design, neural network and data mining, and in particular to a system and method for collecting and importing disease database data based on dynamic forms. Background Art

[0002] Form design technology is a user-centric technology that creates dynamic query forms based on needs. It aims to accurately capture the user's operational intentions, habits, and situational changes, dynamically generate and adjust the form structure, and provide adaptive query forms that are more in line with user needs, ensuring a consistent and smooth user experience in multiple application scenarios. In actual applications, dynamic form design technology analyzes and responds to user input data in real time, and adapts to the content type and context of user input. When users fill out multi-step forms, form design technology will automatically identify and simplify the steps to avoid repeated or erroneous operations. At the same time, it continuously iterates and updates based on user feedback to ensure the accuracy and efficiency of the form.

[0003] Neural network and data mining technology is an advanced data analysis method that combines with each other. It aims to deeply optimize data and extract fine and complex feature information. For data that is rich in content and difficult to merge, neural network technology can achieve effective information association through deep learning of the inherent structure of data, providing a basis for data association and effective information extraction. It is suitable for processing structured data with complex structures and rich information. Neural network and data mining technology can process data patterns and data relationships that are difficult to identify by traditional methods, and provide new solutions for the collection and import of disease-specific database data.

[0004] However, the existing disease-specific database data collection and import systems and methods based on dynamic forms have problems with data formats being diverse, complicated and customized due to the diversity of disease-specific database data fields, numerous types and complex calculation relationships between different data, making data collection and import difficult. In addition, the system development model has problems with untimely response to customized development needs and inflexible system use, which limits the scalability and adaptability of the system. Summary of the invention

[0005] The purpose of the present invention is to provide a system and method for collecting and importing data from a special disease database based on dynamic forms, so as to solve the problems existing in the existing system and method for collecting and importing data from a special disease database based on dynamic forms proposed in the above background technology, namely, the data formats are diversified, complicated and customized, making data collection and import difficult due to the diversity of data fields and types of the special disease database and the complex calculation relationship between different data, and the system development model has the problems of untimely response to customized development needs, inflexible system use and limited system expansion and adaptability.

[0006] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a special disease library data collection and import system and method based on dynamic forms, including a data collection and collation module, a subject library data set creation module, a dynamic form design and management module, a special disease library data configuration management module, a special disease library data retrieval module, and a data visualization module. The data collection and collation module is used to collect user input data, and is characterized in that: the data collection and collation module is used to collect special disease data, and to organize and clean the data to ensure the accuracy and consistency of the data. The data subject library data set creation module is used to create and manage the subject library data set, including defining field names, categories, and codes, to provide basic support for data collection. The dynamic form design and management module includes a dynamic form design unit and a dynamic form management unit. The dynamic form design unit proposes a dynamic form design algorithm based on a probability model, designs and generates dynamic forms according to data characteristics and user needs, and improves the efficiency and adaptability of form design. The dynamic form management unit is used to create a visual management interface and manage dynamic forms. , including the functions of editing, saving, and deleting forms, as well as monitoring and analyzing the use of forms. The special disease library data configuration management module includes a special disease library data conversion unit and a data mining and analysis unit. The data import configuration unit is used to configure and manage the import of data, including converting special disease data in different formats into a unified JSON format, importing content definitions and defining the number of imported content header rows, and specifying import form functions for subsequent data processing and analysis. The data mining and analysis unit proposes a special disease library data mining algorithm based on neural networks, parses the header fields of the imported table, and generates a unique identifier based on the form header structure, and associates it with the previously configured form data structure to ensure the relevance and consistency of the data. The special disease library data retrieval module is used to retrieve data in the special disease library according to user needs, and provides efficient retrieval methods and screening conditions so that users can quickly locate the required special disease library information. The data display and analysis module is used to display and analyze special disease library data, including data visualization, statistical analysis, and report generation functions to help users deeply understand the data.

[0007] Preferably, the data collection and organization module receives data input by the user and organizes it effectively, including standardization of the data format and association between fields, to ensure that the data can be accurately imported into the disease-specific database.

[0008] Preferably, the subject library data set creation module creates and manages the subject library data set by defining the data field names, categories and codes required under different subjects, ensures the standardization and integrity of the subject library data set, and provides basic data format support for data collection.

[0009] Preferably, the dynamic form design and management module includes a dynamic form design unit, which proposes a dynamic form design algorithm based on a probability model, and automatically designs a reasonable dynamic form according to existing data and user needs, including adding calculation rules to realize the calculation relationship between fields to improve the accuracy and efficiency of the design.

[0010] Specifically, the dynamic form design algorithm based on the probability model is as follows: First, according to the created subject library data set, the entities in the subject data set are represented as data objects associated with the user query. The attributes of the entity include name, age, gender, condition and diagnosis result. By selecting the attributes related to the patient entity as the selection component in the dynamic form, the association threshold t is defined to measure the degree of association between the attributes. According to the subject data set architecture, each attribute is traversed and compared with the attributes of the current query entity. If the architecture distance between the attribute in the data set and the attribute of the current query entity is less than or equal to the association threshold t, it is determined that the attribute in the database is associated with the current query entity. The formula for selecting the data set attribute is specifically expressed as:

[0011]

[0012] Among them, H relevant (F) represents the set of attributes associated with the current query entity, F represents the current query form, V represents the attributes of the entity, and H represents the set of all query entity attributes. represents existence, ∈ represents belonging, V j represents the jth attribute of the current query entity, j represents the index of the number of data attributes, H F Represented as the attribute set of the current query entity, d(V, V j ) represents the attribute V and the attribute V j The architectural distance between them is calculated. The attribute of the query in the form is accurately selected through the formula of data set attribute selection. The attribute associated with the current query entity is selected. The association threshold t can adjust the degree of association of the attribute according to the actual situation, making the selection more flexible. Secondly, the best selection component is determined by scoring the entity attributes. The component is represented as an element used to query and display data. In the ranking of the selected component, each attribute corresponds to a component for users to perform query operations. The traditional score TS of the selected component is calculated by precision and recall. The precision calculation formula is expressed as:

[0013]

[0014] Among them, Precision E(F) represents the expected accuracy of the current query form, E represents the expectation, and g represents the data instance in each attribute. Represented as attribute set H F The total number of data instances in , Represented as a user pair data instance level of interest, Represented as data instance The probability of appearing in the subject database dataset, P(σ F |g) indicates that the data instance g satisfies the selection condition σ F The probability of F Represents the conditions used to filter data, N represents the total number of data instances in the subject library dataset, and then the accuracy of the selection component in the dynamic form selection process is measured by the expected precision. The recall calculation formula is expressed as:

[0015]

[0016] Among them, Recall E (F) represents the expected recall rate of the current query form, ξ -1 It is expressed as the proportion of data instances that the user is interested in to the total data instances. The specific calculation formula of the traditional score of the selected component is expressed as:

[0017]

[0018] Among them, Score E (F) represents the traditional score of the selected component in the form of the current query, β represents the weighting factor, which is used to balance the relationship between precision and recall. The probability model is constructed based on the relevance of the attributes in the entity and the historical click data of the selected component. The specific formula is expressed as follows:

[0019] P model (F) = (1-λ)·P history (s|H F )+λ·P schema (V j |H F )

[0020] Among them, P model (F) represents the probability model score of the selected component under the current query form, P history (s|H F ) is represented by the historical click data s based on the selection component and the current query entity attribute set H F The conditional probability, P schema (V j |H F) is represented by the jth attribute V in the attributes of the current query entity j With the current query entity attribute set H F The conditional probability of , s represents the historical click data of the selection component, λ represents the weight parameter used to balance the relevance of the attributes in the entity and the contribution of the historical click data of the selection component. The constructed probability model evaluates the quality of the selection component according to the historical data and the relevance of the attributes, thereby improving the efficiency and accuracy of the user query form. Finally, the selection components are ranked by the constructed probability model and the traditional score TS of the selection component, and the priority order of the selection component of each attribute is determined, so that users can efficiently obtain the most relevant and useful selection components when filling in the form. The ranking score of the selection component is calculated according to the following formula:

[0021] TS=α·Score E (F)+(1-α)·P model (F)

[0022] Among them, TS represents the ranking score of the selection component, α represents the parameter that weighs the traditional score and the probability model score. By comprehensively considering the traditional score and the probability model score, the priority order of the selection component for each attribute can be determined, which can more accurately reflect the importance and relevance of the selection component, so that users can efficiently obtain the most relevant and useful selection components.

[0023] Preferably, the dynamic form design and management module includes a dynamic form management unit, which realizes visual control of the data collection page in a visual manner by dragging and dropping. The user can quickly design and adjust the form layout through simple operations to meet different data collection scenarios and ensure the accuracy of data collection.

[0024] Preferably, the special disease database data configuration management module includes a data import configuration unit and a data mining and analysis unit. The data import configuration unit ensures that special disease data in different formats are converted into a unified JSON format through data conversion operations to facilitate subsequent processing and analysis, and defines the number of header rows of the imported content and the corresponding import form to ensure that the imported data and the values ​​of the dynamic form design can be correctly mapped and unified.

[0025] Preferably, the special disease database data configuration management module includes a data mining and analysis unit, which proposes a special disease database data mining algorithm based on a neural network, analyzes and mines the imported special disease data, discovers the correlation in the data, parses the header fields of the imported table, and generates a unique identifier based on the form header structure to achieve data consistency, ensure that the imported data matches the existing data structure in the system, and avoid data redundancy and inconsistency.

[0026] Specifically, the neural network-based disease database data mining algorithm is as follows: First, a neural network model is constructed to mine hidden information in the structured disease database data after the unified format to find the correlation between the data. The constructed neural network model is as follows:

[0027] K=(L i , P j )

[0028] Among them, K represents the neural network model for classifying hidden information, L i A non-empty finite set of neural network nodes, P j Represented as a finite set of network edges, in a neural network, data is represented as a set of nodes L i and edge set P j The directed neural network model K is constructed. The nodes are represented as data, including various cases and disease parameters in the disease database. The edges represent the relationship between nodes, which represents the similarity and correlation between data. The complex relationship between data is captured by neural network in order to understand the correlation structure between data. Secondly, the information classification model is constructed by neural network, and the information classification standard is established to realize the mining and correlation analysis of hidden information in the disease database. The feature vector representation of information is introduced, and the neural network is used for training and optimization to obtain more accurate information correlation weights. The specific formula of the information classification model is expressed as follows:

[0029]

[0030] Where π(X) represents the standard information classification model of the degree of association between data, X represents the information sample set, δ represents the information association weight, Δ(X′) represents the feature vector of the information sample, X′ represents the information sample set to be mined, N represents the number of information samples, n represents the index of the number of information samples, M represents the number of features of the information sample, m represents the index of the number of features of the information sample, and x m Represented as the mth eigenvalue of the information sample, It is represented by the normalized weight value of the association between the mth feature of the nth information sample and the overall information after optimization by the neural network K, ξ represents the neural network node, ψ represents the network edge, and p n It is expressed as the probability density of the nth information sample. By constructing a standard model for information classification, the degree of association between information can be determined, providing a basis for subsequent data mining. Then, the information association normalization weight value ω(ξ, ψ) is calculated to determine the degree of association between hidden information. The specific calculation formula is expressed as:

[0031]

[0032] Among them, ω D (Y) represents the weight value of the label D corresponding to ξ, ψ in the newly added information Y, ω D (X) represents the weight value of the label D corresponding to ξ, ψ in the existing information X, J(ω, b) represents the loss function of the neural network model, α represents the learning rate of the neural network model, It is expressed as a partial derivative operation to calculate the probability density p of the information sample. The specific calculation formula is:

[0033] p=p(Y,λ)·p(X,λ)

[0034] Among them, p(X, λ) represents the probability density of the existing information X, p(Y, λ) represents the probability density of the newly added information Y, and λ represents the regularization parameter. Finally, by calculating the loss function of the neural network model, the prediction effect of the model on the correlation of hidden information in the disease database is measured, and then the neural network model is optimized. The calculation formula of the loss function J(ω) is expressed as:

[0035]

[0036] Among them, π(X (n) ) is represented as the nth sample X (n) The degree of information association with the sample set X, The predicted value of the nth sample, b is represented as a bias term, and the neural network model prediction formula is expressed as:

[0037]

[0038] Among them, σ represents the activation function. The optimized neural network model is used to predict the data in the disease database to obtain the predicted correlation results of hidden information. By analyzing the prediction results, the correlation relationship between the data is mined to discover the potential information correlation rules.

[0039] Preferably, the disease-specific database data retrieval module retrieves and queries the data in the disease-specific database by providing efficient retrieval conditions and flexible query methods, ensuring that users can find the required data quickly and accurately to meet the needs of different users.

[0040] Preferably, the data display and analysis module visualizes the data in the disease-specific database, generates analysis reports and real-time data updates, thereby ensuring that users can intuitively and deeply understand the data in the disease-specific database and helping users to use the data efficiently.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] 1. The dynamic form design unit proposes a dynamic form design algorithm based on a probability model, aiming to achieve a design method for dynamic forms that can be efficiently dynamically adjusted and scalable. First, based on the created subject library dataset, the algorithm represents the entities in the subject dataset as data objects associated with the user query, whose attributes include name, age, gender, condition, and diagnosis results. It is user-centric and uses implicit user feedback to estimate the user's interests and preferences, selects attributes related to the patient entity as the selection component in the dynamic form, and defines an association threshold to measure the degree of association between attributes to accurately select the attributes queried in the form and the attributes associated with the current query entity. Secondly, the expected performance of the query form is evaluated using the expected precision and expected recall rate indicators to determine the best selection component based on the score of the entity attributes, and the traditional score of the selection component is obtained by calculating the precision and recall rate. The precision calculation formula measures the selection component in the dynamic form based on the expected precision of the current query form. The accuracy of the dynamic form selection process is calculated, and the recall calculation formula evaluates the proportion of data instances that users are interested in to the total data instances. Then, a probability model is constructed based on the relevance of attributes in the entity and the historical click data of the selection component to evaluate the quality of the selection component. The probability model score comprehensively considers the conditional probability of the historical click data of the selection component and the current query entity attribute set, as well as the weight parameter that balances the relevance of the attributes in the entity and the contribution of the historical click data of the selection component. Finally, the selection components are ranked by the constructed probability model and the traditional score of the selection component to determine the priority order of the selection component for each attribute, which can more accurately reflect the importance and relevance of the selection component, so that users can efficiently obtain the most relevant and useful selection components. In summary, the dynamic form design algorithm based on the probability model proposed by the dynamic form design unit provides an effective way to achieve an efficient, dynamically adjusted and scalable dynamic form design method.

[0043] 2. The data mining and analysis unit proposes a data mining algorithm for the disease-specific database based on neural networks. First, the algorithm understands and captures the complex relationships between data by constructing a neural network model, and represents the data as a directed neural network model of a node set and an edge set, where the nodes represent the data, including various cases and disease parameters in the disease-specific database, and the edges represent the relationships between nodes, that is, the similarities and correlations between the data. The constructed neural network model can effectively mine the hidden information between the data, and provide a basis for subsequent data collection and import. Secondly, the neural network is used to construct an information classification model, and an information classification standard model is established, so as to realize the mining and correlation analysis of hidden information in the disease-specific database. By introducing the feature vector representation of the information and using the neural network for training and optimization, the algorithm obtains more accurate information correlation weights, so that the degree of correlation between the data can be accurately determined, which provides a basis for subsequent Then, the information association normalization weight value is calculated according to the information association classification model to determine the degree of association between hidden information. The neural network model is optimized through the learning rate and loss function of the neural network model, thereby improving the model's prediction effect on the correlation of hidden information in the special disease database. Finally, through the prediction results of the neural network model, the correlation relationship between the data is mined to discover the potential information association rules. By analyzing the prediction results, we can have a deeper understanding of the hidden information in the special disease database data, and provide valuable reference for further research and application in the medical field. In summary, the dynamic form design algorithm based on the probability model realizes the hidden information mining and association analysis of the special disease database data through the design and optimization of the neural network model, which provides an effective solution for the collection and import of special disease database data in the medical field, and has important practical value and application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The invention is further illustrated by the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0045] Figure 1 It is a schematic diagram of the structure of the present invention. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] See also Figure 1The present invention provides a disease-specific database data collection and import system and method based on dynamic forms, including a data collection and organization module, a subject database data set creation module, a dynamic form design and management module, a disease-specific database data configuration management module, a disease-specific database data retrieval module, and a data visualization module. The data collection and organization module is used to collect user input data, and is characterized in that: the data collection and organization module is used to collect disease-specific data, and organize and clean the data to ensure the accuracy and consistency of the data; the subject database data set creation module is used to create and manage subject database data sets, including defining field names, categories, and codes, to provide basic support for data collection; the dynamic form design and management module includes a dynamic form design unit and a dynamic form management unit; the dynamic form design unit proposes a dynamic form design algorithm based on a probability model, designs and generates dynamic forms according to data characteristics and user needs, and improves the efficiency and adaptability of form design; the dynamic form management unit is used to create a visual management interface, manage dynamic forms, including the editing of forms The data mining and analysis unit is used to configure and manage the import of data, including converting special disease data in different formats into a unified JSON format, importing content definition and defining the number of import content header rows, and specifying the import form function for subsequent data processing and analysis. The data mining and analysis unit proposes a special disease database data mining algorithm based on neural networks, parses the header fields of the imported table, generates a unique identifier based on the form header structure, and associates it with the previously configured form data structure to ensure the relevance and consistency of the data. The special disease database data retrieval module is used to retrieve data in the special disease database according to user needs, and provides efficient retrieval methods and screening conditions so that users can quickly locate the required special disease database information. The data display and analysis module is used to display and analyze the special disease database data, including data visualization, statistical analysis, and report generation functions to help users deeply understand the data.

[0048] See also Figure 1 ,Furthermore, the data collection and organization module receives the data input by the ,users and organizes it effectively, including the standardization of the ,data format and the association between the fields, to ensure that the data can be ,accurately imported into the disease database.

[0049] See also Figure 1 ,Furthermore, the theme library dataset creation module creates and manages the theme library dataset by defining ,the data field names, categories and codes required under different ,themes, ensuring the standardization and integrity of the theme library dataset and ,providing basic data format support for data collection.

[0050] See also Figure 1,Furthermore, the dynamic form design and management module includes a ,dynamic form design unit, which proposes a dynamic form design algorithm based on ,probabilistic model, and automatically designs reasonable dynamic forms according to ,existing data and user needs, including adding calculation rules and ,realizing the calculation relationship between fields to improve the ,accuracy and efficiency of the design.

[0051] See also Figure 1 ,Furthermore, the dynamic form design algorithm based on the probability model is as follows: ,First, according to the created subject library dataset, the entities in the subject dataset are represented as ,data objects associated with the user query. The attributes of the entity include name, ,age, gender, condition, and diagnosis results. ,By selecting the attributes related to the patient entity as the ,selection components in the dynamic form, the association threshold t is defined to ,measure the degree of association between the attributes. According to the ,subject dataset architecture, each attribute is traversed and ,compared with the attributes of the current query entity. If the ,architecture distance between the attribute in the dataset and the attribute of the ,current query entity is less than or equal to the association threshold t, ,then it is determined that the attribute in the database is ,associated with the current query entity. The formula for selecting the dataset ,attributes is specifically expressed as:

[0052]

[0053] Among them, H relevant (F) represents the set of attributes associated with the current query entity, F represents the current query form, V represents the attributes of the entity, and H represents the set of all query entity attributes. represents existence, ∈ represents belonging, V j represents the jth attribute of the current query entity, j represents the index of the number of data attributes, H F Represented as the attribute set of the current query entity, d(V, V j ) represents the attribute V and the attribute V j The architectural distance between them is calculated. The attribute of the query in the form is accurately selected through the formula of data set attribute selection. The attribute associated with the current query entity is selected. The association threshold t can adjust the degree of association of the attribute according to the actual situation, making the selection more flexible. Secondly, the best selection component is determined by scoring the entity attributes. The component is represented as an element used to query and display data. In the ranking of the selected component, each attribute corresponds to a component for users to perform query operations. The traditional score TS of the selected component is calculated by precision and recall. The precision calculation formula is expressed as:

[0054]

[0055] Among them, Precision E(F) represents the expected accuracy of the current query form, E represents the expectation, and g represents the data instance in each attribute. Represented as attribute set H F The total number of data instances in , Represented as a user pair data instance level of interest, Represented as data instance The probability of appearing in the subject database dataset, P(σ F |g) indicates that the data instance g satisfies the selection condition σ F The probability of F Represents the conditions used to filter data, N represents the total number of data instances in the subject library dataset, and then the accuracy of the selection component in the dynamic form selection process is measured by the expected precision. The recall calculation formula is expressed as:

[0056]

[0057] Among them, Recall E (F) represents the expected recall rate of the current query form, ξ -1 It is expressed as the proportion of data instances that the user is interested in to the total data instances. The specific calculation formula of the traditional score of the selected component is expressed as:

[0058]

[0059] Among them, Score E (F) represents the traditional score of the selected component in the form of the current query, β represents the weighting factor, which is used to balance the relationship between precision and recall. The probability model is constructed based on the relevance of the attributes in the entity and the historical click data of the selected component. The specific formula is expressed as follows:

[0060] P model (F) = (1-λ)·P history (s|H F )+λ·P schema (V j |H F )

[0061] Among them, P model (F) represents the probability model score of the selected component under the current query form, P history (s|H F ) is represented by the historical click data s based on the selection component and the current query entity attribute set H F The conditional probability, P schema (V j |H F) is represented by the jth attribute V in the attributes of the current query entity j With the current query entity attribute set H F The conditional probability of , s represents the historical click data of the selection component, λ represents the weight parameter used to balance the relevance of the attributes in the entity and the contribution of the historical click data of the selection component. The constructed probability model evaluates the quality of the selection component according to the historical data and the relevance of the attributes, thereby improving the efficiency and accuracy of the user query form. Finally, the selection components are ranked by the constructed probability model and the traditional score TS of the selection component, and the priority order of the selection component of each attribute is determined, so that users can efficiently obtain the most relevant and useful selection components when filling in the form. The ranking score of the selection component is calculated according to the following formula:

[0062] TS=α·Score E (F)+(1-α)·P model (F)

[0063] Among them, TS represents the ranking score of the selection component, α represents the parameter that weighs the traditional score and the probability model score. By comprehensively considering the traditional score and the probability model score, the priority order of the selection component for each attribute can be determined, which can more accurately reflect the importance and relevance of the selection component, so that users can efficiently obtain the most relevant and useful selection components.

[0064] See also Figure 1 ,Furthermore, the dynamic form design and management module includes a ,dynamic form management unit, which realizes visual control of the data ,collection page in a visual way by dragging and dropping. Users can ,quickly design and adjust the form layout through simple ,operations to meet different data collection scenarios and ensure the ,accuracy of data collection.

[0065] See also Figure 1 Furthermore, the special disease database data configuration management module includes a data import configuration unit and a data mining and analysis unit. The data import configuration unit ensures that special disease data in different formats are converted into a unified JSON format through data conversion operations to facilitate subsequent processing and analysis, and defines the number of header rows of the imported content and the corresponding import form to ensure that the imported data and the values ​​of the dynamic form design can be correctly mapped and unified.

[0066] See also Figure 1Furthermore, the special disease database data configuration management module includes a data mining and analysis unit, which proposes a special disease database data mining algorithm based on neural network, analyzes and mines the imported special disease data, discovers the correlation in the data, parses the header fields of the imported table, and generates a unique identifier based on the form header structure to achieve data consistency, ensure that the imported data matches the existing data structure in the system, and avoid data redundancy and inconsistency.

[0067] See also Figure 1 ,Furthermore, the data mining algorithm of the disease-specific database based on neural network is as follows: First, a neural network model is constructed to mine the hidden information in the structured disease-specific database data after the unified format to find the correlation between the data. The constructed neural network model is as follows:

[0068] K=(L i , P j )

[0069] Among them, K represents the neural network model for classifying hidden information, L i A non-empty finite set of neural network nodes, P j Represented as a finite set of network edges, in a neural network, data is represented as a set of nodes L i and edge set P j The directed neural network model K is constructed. The nodes are represented as data, including various cases and disease parameters in the disease database. The edges represent the relationship between nodes, which represents the similarity and correlation between data. The complex relationship between data is captured by neural network in order to understand the correlation structure between data. Secondly, the information classification model is constructed by neural network, and the information classification standard is established to realize the mining and correlation analysis of hidden information in the disease database. The feature vector representation of information is introduced, and the neural network is used for training and optimization to obtain more accurate information correlation weights. The specific formula of the information classification model is expressed as follows:

[0070]

[0071] Where π(X) represents the standard information classification model of the degree of association between data, X represents the information sample set, δ represents the information association weight, Δ(X′) represents the feature vector of the information sample, X′ represents the information sample set to be mined, N represents the number of information samples, n represents the index of the number of information samples, M represents the number of features of the information sample, m represents the index of the number of features of the information sample, and x m Represented as the mth eigenvalue of the information sample, It is represented by the normalized weight value of the association between the mth feature of the nth information sample and the overall information after optimization by the neural network K, ξ represents the neural network node, ψ represents the network edge, and p n It is expressed as the probability density of the nth information sample. By constructing a standard model for information classification, the degree of association between information can be determined, providing a basis for subsequent data mining. Then, the information association normalization weight value ω(ξ, ψ) is calculated to determine the degree of association between hidden information. The specific calculation formula is expressed as:

[0072]

[0073] Among them, ω D (Y) represents the weight value of the label D corresponding to ξ, ψ in the newly added information Y, ω D (X) represents the weight value of the label D corresponding to ξ, ψ in the existing information X, J(ω, b) represents the loss function of the neural network model, α represents the learning rate of the neural network model, It is expressed as a partial derivative operation to calculate the probability density p of the information sample. The specific calculation formula is:

[0074] p=p(Y,λ)·p(X,λ)

[0075] Among them, p(X, λ) represents the probability density of the existing information X, p(Y, λ) represents the probability density of the newly added information Y, and λ represents the regularization parameter. Finally, by calculating the loss function of the neural network model, the prediction effect of the model on the correlation of hidden information in the disease database is measured, and then the neural network model is optimized. The calculation formula of the loss function J(ω) is expressed as:

[0076]

[0077] Among them, π(X (n) ) is represented as the nth sample X (n) The degree of information association with the sample set X, The predicted value of the nth sample, b is represented as a bias term, and the neural network model prediction formula is expressed as:

[0078]

[0079] Among them, σ represents the activation function. The optimized neural network model is used to predict the data in the disease database to obtain the predicted correlation results of hidden information. By analyzing the prediction results, the correlation relationship between the data is mined to discover the potential information correlation rules.

[0080] See also Figure 1,Furthermore, the data retrieval module of the special disease database retrieves and queries the data in the ,special disease database by providing efficient retrieval conditions and flexible ,query methods, ensuring that users can find the required data quickly and accurately ,to meet the needs of different users.

[0081] See also Figure 1 ,Furthermore, the data display and analysis module visualizes the data in the ,special disease database, and generates analysis reports and real-time data updates, ,ensuring that users can intuitively and deeply understand the data in the ,special disease database, and helping users to use the data efficiently.

[0082] When used specifically, first, the data collection and organization module collects disease-specific data, and organizes and cleans the data to ensure the accuracy and consistency of the data. The data base dataset creation module creates and manages the data base dataset, including defining field names, categories, and codes, to provide basic support for data collection. Then, the dynamic form design and management module includes a dynamic form design unit and a dynamic form management unit. The dynamic form design unit proposes a dynamic form design algorithm based on a probability model, designs and generates dynamic forms according to data characteristics and user needs, and improves the efficiency and adaptability of form design. The dynamic form management unit creates a visual management interface to manage dynamic forms, including form editing, saving, and deletion functions, as well as monitoring and analysis of form usage. Secondly, the disease-specific database data configuration management module includes a disease-specific database data conversion unit and a data mining and analysis unit. The data import configuration unit configures and manages the import of data, including converting special disease data in different formats into a unified JSON format, importing content definitions and defining the number of import content header rows, and specifying import form functions for subsequent data processing and analysis. The data mining and analysis unit proposes a special disease database data mining algorithm based on neural networks, parses the header fields of the imported table, generates a unique identifier based on the form header structure, and associates it with the previously configured form data structure to ensure the relevance and consistency of the data. Then, the special disease database data retrieval module retrieves data in the special disease database according to user needs, and provides efficient retrieval methods and filtering conditions so that users can quickly locate the required special disease database information. Finally, the data display and analysis module is used to display and analyze special disease database data, including data visualization, statistical analysis, and report generation functions to help users deeply understand the data.

[0083] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A disease database data collection and import system based on dynamic forms, including a data collection and sorting module, a subject database data set creation module, a dynamic form design and management module, a disease database data configuration management module, a disease database data retrieval module, and a data visualization module, characterized in that: The data collection and organization module is used to collect disease-specific data, and organize and clean the data to ensure the accuracy and consistency of the data. The subject library data set creation module is used to create and manage subject library data sets, including defining field names, categories, and codes to provide basic support for data collection. The dynamic form design and management module includes a dynamic form design unit and a dynamic form management unit. The dynamic form design unit proposes a dynamic form design algorithm based on a probability model, designs and generates dynamic forms according to data characteristics and user needs, and improves the efficiency and adaptability of form design. The dynamic form management unit is used to create a visual management interface and manage dynamic forms, including form editing, saving, and deletion functions, as well as monitoring and analysis of form usage. The disease-specific library data configuration management module includes a data import configuration unit and a data mining analysis unit. The import configuration unit is used to configure and manage the import of data, including converting disease-specific data in different formats into a unified JSON format, importing content definitions and defining the number of import content header rows, and specifying import form functions for subsequent data processing and analysis. The data mining and analysis unit proposes a disease-specific database data mining algorithm based on a neural network, parses the header fields of the imported form, generates a unique identifier based on the form header structure, and associates it with the previously configured form data structure to ensure the relevance and consistency of the data. The disease-specific database data retrieval module is used to retrieve data in the disease-specific database according to user needs, and provides efficient retrieval methods and screening conditions so that users can quickly locate the required disease-specific database information. The data visualization module is used to display and analyze disease-specific database data, including data visualization, statistical analysis, and report generation functions to help users deeply understand the data. The dynamic form design and management module includes a dynamic form design unit and a dynamic form management unit. The dynamic form design unit proposes a dynamic form design algorithm based on a probability model, and automatically designs a reasonable dynamic form according to existing data and user needs, including adding calculation rules to realize the calculation relationship between fields; The dynamic form management unit realizes visual control of the data collection page through visual means and drag-and-drop; According to the created subject database dataset, the entities in the subject dataset are represented as data objects associated with the user query. The attributes of the entity include name, age, gender, condition, and diagnosis result. By selecting the attributes related to the patient entity as the selection component in the dynamic form, the association threshold t is defined to measure the degree of association between the attributes. The formula for selecting the dataset attributes is specifically expressed as follows: Among them, H relevant (F) represents the set of attributes associated with the current query entity, F represents the current query form, V represents the attributes of the entity, and H represents the set of all query entity attributes. represents existence, ∈ represents belonging, V j represents the jth attribute of the current query entity, j represents the index of the number of data attributes, H F Represented as the attribute set of the current query entity, d(V,V j ) represents the attribute V and the attribute V j The architectural distance between entities is determined by scoring entity attributes. Components are represented as elements used to query and display data. The traditional score TS of the selected component is calculated by precision and recall. The specific calculation formula of the traditional score of the selected component is expressed as: Among them, Score E (F) represents the traditional score of the selected component in the form of the current query, β represents the weighting factor, and Precision E (F) represents the expected accuracy of the current query form, and E represents the expected, Recall E (F) represents the expected recall rate of the form of the current query, which is used to balance the relationship between precision and recall rate. The probability model is constructed based on the relevance of the attributes in the entity and the historical click data of the selected component. The specific formula is expressed as: P model (F)=(1-λ)·P history (s|H F )+λ·P schema (V j |H F ) Among them, P model (F) represents the probability model score of the selected component under the current query form, P history (s|H F ) is represented by the historical click data s based on the selection component and the current query entity attribute set H F The conditional probability, P schema (V j |H F ) is represented by the jth attribute V in the attributes of the current query entity j With the current query entity attribute set H F The conditional probability of , s represents the historical click data of the selection component, and λ represents the weight parameter used to balance the relevance of the attribute in the entity and the contribution of the historical click data of the selection component. The selection components are ranked by the constructed probability model and the traditional score TS of the selection component to determine the priority order of the selection component for each attribute, so that users can efficiently obtain the most relevant and useful selection components when filling out the form. The ranking score of the selection component is calculated according to the following formula: TS=α·Score E (F)+(1-α)·P model (F) Among them, TS represents the ranking score of the selection component, and α represents the parameter that weighs the traditional score and the probability model score. By comprehensively considering the traditional score and the probability model score and then determining the priority order of the selection component for each attribute, it can more accurately reflect the importance and relevance of the selection component, so that users can efficiently obtain the most relevant and useful selection components.

2. According to claim 1, a dynamic form-based disease database data collection and import system is characterized by: The data collection and organization module receives the data input by the user and organizes it effectively, including standardization of the data format and association between fields, to ensure that the data can be accurately imported into the disease-specific database.

3. According to claim 1, a dynamic form-based disease database data collection and import system is characterized by: The subject database dataset creation module creates and manages the subject database dataset by defining the data field names, categories and codes required under different subjects, ensures the standardization and integrity of the subject database dataset, and provides basic data format support for data collection.

4. According to the dynamic form-based disease database data collection and import system of claim 1, it is characterized in that: The disease-specific database data configuration management module includes a data import configuration unit and a data mining and analysis unit. The data import configuration unit ensures that disease-specific data in different formats are converted into a unified JSON format through data conversion operations to facilitate subsequent processing and analysis, and defines the number of header rows of the imported content and the corresponding import form to ensure that the imported data and the values ​​of the dynamic form design can be correctly mapped and unified; The data mining and analysis unit proposes a data mining algorithm for the disease-specific database based on a neural network. By analyzing and mining the imported disease-specific data, it discovers the correlation in the data, parses the header fields of the imported tables, and generates a unique identifier based on the form header structure to achieve data consistency and ensure that the imported data matches the existing data structure in the system to avoid data redundancy and inconsistency.

5. According to claim 4, a dynamic form-based disease database data collection and import system is characterized by: First, a neural network model is constructed to mine the hidden information in the structured disease database data with unified format to find the correlation between the data. The constructed neural network model is shown below: K=(L i ,P j ) Among them, K represents the neural network model for classifying hidden information, L i A non-empty finite set of neural network nodes, P j Represented as a finite set of network edges, in a neural network, data is represented as a set of nodes L i and edge set P j The directed neural network model K, the nodes are represented as data, and the specific formula of the information classification model is expressed as: Among them, π(X) represents the standard information classification model of the degree of association between data, X represents the information sample set, δ represents the information association weight, Δ(X′) represents the feature vector of the information sample, X′ represents the information sample set to be mined, N represents the number of information samples, n represents the index of the number of information samples, M represents the number of features of the information sample, m represents the index of the number of features of the information sample, and x m Represented as the mth eigenvalue of the information sample, It is represented by the normalized weight value of the association between the mth feature of the nth information sample and the overall information after optimization by the neural network K, ξ represents the neural network node, ψ represents the network edge, and p n Expressed as the probability density of the nth information sample, calculate the information association normalized weight value ω (ξ，ψ) , the specific calculation formula is expressed as: Among them, ω D (Y) represents the weight value of the label D corresponding to ξ, ψ in the newly added information Y, ω D (X) represents the weight value of the label D corresponding to ξ, ψ in the existing information X, J(ω, b) represents the loss function of the neural network model, α represents the learning rate of the neural network model, It is expressed as a partial derivative operation to calculate the probability density p of the information sample. The specific calculation formula is expressed as: p=p(Y,λ)·p(X,λ) Among them, p(X, λ) represents the probability density of the existing information X, p(Y, λ) represents the probability density of the newly added information Y, λ represents the regularization parameter, and the loss function J(ω, b) is calculated as follows: Among them, π(X (n) ) is represented as the nth sample X (n) The degree of information association with the sample set X, The predicted value of the nth sample, b is represented as a bias term, and the neural network model prediction formula is expressed as: Among them, σ represents the activation function. The optimized neural network model is used to predict the data in the disease database to obtain the predicted correlation results of hidden information. By analyzing the prediction results, the correlation relationship between the data is mined to discover the potential information correlation rules.

6. According to the dynamic form-based disease database data collection and import system of claim 1, it is characterized in that: The disease-specific database data retrieval module retrieves and queries the data in the disease-specific database by providing efficient retrieval conditions and flexible query methods, ensuring that users can find the required data quickly and accurately to meet the needs of different users.

7. According to claim 1, a dynamic form-based disease database data collection and import system is characterized by: The data visualization module visualizes the data in the disease-specific database and generates analysis reports and real-time data updates, ensuring that users can intuitively and deeply understand the data in the disease-specific database and helping users use the data efficiently.

Images