Big data severe pneumonia data classification and automatic screening method

Abstract

The invention relates to the field of medical data classification, in particular to a big data severe pneumonia data classification and automatic screening method, a medical algorithm model is constructed, and a clinical diagnosis and treatment thinking logic process is restored. After the microbiological detection data identification degree of the microbiological detection pollution interference elimination model is enhanced, a multi-modal fusion technology is adopted, and finally, multi-task output of cause diagnosis and early warning of severe pneumonia is formed; aiming at insufficient feature processing, insufficient medical logic consideration and improper fusion strategy, high identification of microorganism features and dual identification of microorganisms and host factors are carried out, a feature level is combined with modal identification, and completeness, specificity and logicality are better achieved through multi-layer processing. Diagnosis of severe pneumonia pathogens is accurate according to neural network learning, and data are classified according to diagnosis results and severe pneumonia weights.

Description

Technical Field

[0001] The present application relates to the field of medical data classification, and in particular to a method for classifying and automatically screening big data severe pneumonia data. Background Art

[0002] There are still many problems in the application of precise etiological diagnosis, such as the difficulty in obtaining deep respiratory tract samples, long detection time, poor differentiation between pathogenic microorganisms and colonizing microorganisms, background microorganisms and contaminating microorganisms, lack of unified judgment criteria for interpreting results, and low detection rate of some infectious pathogens. Delays in etiological diagnosis can easily lead to inappropriate treatment and increase mortality. At the same time, the widespread use of broad-spectrum antibiotics can easily lead to drug resistance in pathogens and increase side effects. In summary, the existing diagnostic methods based on microbial detection are still difficult to achieve the clinical requirements of early, rapid and accurate diagnosis of severe pneumonia. Finding more efficient and accurate early etiological diagnosis methods is an urgent clinical need, and there is an urgent need to improve the accuracy of lung disease data collection. Summary of the Invention

[0003] The purpose of the present invention is to propose a method for classifying and automatically screening big data severe pneumonia data to solve one or more technical problems existing in the prior art and at least provide a beneficial option or create conditions.

[0004] A big data classification and automated screening method for severe pneumonia uses standardized processes to collect and process data across different dimensionalities. Data collection covers patient baseline information, laboratory test data, imaging data, interventions, and clinical outcomes. Laboratory test data will be uniformly converted to international standard units, and imaging data will be stored in DICOM format to ensure compatibility across hospital systems. Metadata will be appended to all data collection processes.

[0005] In terms of data processing, all multimodal data will be cleaned and standardized to ensure data consistency, completeness and accuracy. All patients' clinical data will be recorded using standardized forms, including scores on the day of admission, including PSI, APACHE II, CURB-65 and SOFA scores, as well as dynamic changes in the disease. The data cleaning process includes checking missing values, removing duplicate data and processing outliers to ensure the quality of the data finally included in the database is reliable.

[0006] The database architecture will adopt a hybrid model, combining the strengths of relational and NoSQL databases. For structured data, relational databases such as MySQL or PostgreSQL will be used to ensure consistent and scalable data storage. For unstructured data, NoSQL databases such as MongoDB will be used to support parallel processing and distributed storage of massive amounts of data, enhancing flexibility in data expansion and retrieval.

[0007] The database will utilize a distributed storage structure to ensure data security and efficient management, supporting simultaneous access by multiple users and hospitals. The system provides a RESTful API interface to facilitate data access and sharing with external systems, enabling flexible data invocation and integration. Regarding data permission management, the system will set different access levels based on user roles to ensure the privacy of sensitive data in compliance with international privacy regulations. Patients' personal privacy information will be anonymized, and encryption technology will be used to ensure data security during storage and transmission.

[0008] Furthermore, based on the two core diagnostic data of microbial detection and host information, multi-layer simulation diagnosis is performed and a training model is output, including the following steps:

[0009] Establish the weights of two core diagnostic data: microbial detection and host information;

[0010] By comparing the weights of the file contributions in the forward and reverse weight directions, the neural network input is compared with the weights of the file contributions. The network output is determined by the forward and reverse outputs. The calculation formula is as follows:

[0011]

[0012] O t =g(Vh t +b3);

[0013] O t =g(Vh t +b3);

[0014] Wherein, x represents the input value, which is the two core diagnostic data of microbial detection and host information. The weight values ​​of the two core diagnostic data of microbial detection and host information are defined based on the multimodal fusion model. are the forward weight and reverse weight directions respectively, is the network forward output value, W1, U1, b1 are the input layer weight, hidden layer weight and bias vector of the forward output respectively, is the reverse output value of the network, W2, U2, and b2 are the input layer weight, hidden layer weight, and bias vector of the reverse output respectively. The input layer weight is the result obtained by comparing the forward output and the reverse output. L1 is the input layer weight, L2 is the hidden layer weight, and the bias vector is and If the comparison result is L2, the input layer weight, L1, the hidden layer weight, and the bias vector is and of and;

[0015] O t is the final output value, V and b3 are the output layer weight matrix and bias vector respectively, and the symbol is the splicing operation;

[0016] The final output value O t The neural network model is put into training. After the training is completed, the input data is quickly output as an indicator X. The indicator X is used to determine whether the comparison data meets the qualified standard. Each indicator X needs to be trained through a separate neural network model to determine the mapping relationship between the input data and the indicator. The files are automatically classified through the output of the neural network model and the results of the training output are output multiple times.

[0017] Furthermore, in the neural network model, the number of neurons in the network input layer is equal to the number of input data types, the number of neurons in the output layer is equal to the number of output data types, and the number of intermediate layers and the number of neurons in each layer are freely set. The setting of the intermediate layers will directly affect the accuracy of the approximate data of the neural network. Whether the number of intermediate layers or the number of neurons needs to be increased or decreased is determined based on the training results. The calculation method for determining the number of intermediate layers and the number of neurons is:

[0018] The middle layers in the neural network model are added layer by layer, and the output index X corresponding to the number of middle layers is named. When the number of middle layers is 1, the output index X1 is output, and so on. Based on the index X, a nonlinear regression function is constructed. a is a constant value, and the variance D of the function's independent variable is calculated. Where L is the number of intermediate layers at output, X i is the output index X, the X i ` is the input data corresponding to the output index X, when When A takes the minimum value, To obtain the partial derivative of the function, A is a parameter that determines the number of intermediate layers. The value of A is compared with the variance D. If D>A, the output index X is input into the neural network model, and the number of intermediate layers is increased by 1. If D≤A, the output index X is output.

[0019] Update the weight coefficients of neurons using the gradient descent method:

[0020]

[0021] in, To obtain the partial derivative of the function, P and Q are the system output error and the neuron weight increment, respectively, both of which are constant values, ω(l) is the neuron weight coefficient, Δω(l) is the updated neuron weight coefficient, δ represents the neuron learning rate, and the number of neurons set in the middle layer is determined by the Δω(l). According to the superposition analysis algorithm combined with the output value O t The algorithm is deeply learned to output top-level comparative analysis data and second-level comparative analysis data to ensure the quality of the comparative data.

[0022] Further, automatic classification is performed based on the weight of the data quality and severe pneumonia data, a memory coupled to the one or more processors, the memory storing instructions that, when executed by the one or more processors, cause the one or more processors to: provide a classification pipeline comprising a component that obtains metadata associated with a data item and existing classification metadata associated with the data item, wherein the existing classification metadata comprises a current classification value of the data item, and wherein the current classification value of the data item is stored in the data item;

[0023] A plurality of classifier modules are provided, wherein each of the plurality of classifier modules has an associated classification rule, and wherein each of the classification rules, when invoked, uses metadata associated with a data item and existing classification metadata associated with the data item to determine classification metadata for the data item, and wherein the plurality of classifier modules are mediated using at least one of: aggregate classification, authoritative classification, and high classification.

[0024] Furthermore, the method is based on a system, which includes: a processor and a memory. The processor and memory in the overall system can both run a computer program in the processor. When the processor executes the computer program, it can implement the steps of a big data severe pneumonia data classification and automatic screening method as described in any one of claims 1-4.

[0025] The beneficial effects of the present invention are: in response to insufficient feature processing, insufficient medical logic considerations and inappropriate fusion strategies, high-level identification of microbial features, dual identification of microorganisms and host factors, and combination of feature levels with modal identification are performed. After multi-layer processing, it is more complete, specific and logical, and the diagnosis of pathogens of severe pneumonia is accurately made based on neural network learning, and the data is classified according to the diagnosis results and severe pneumonia weights. DETAILED DESCRIPTION

[0026] In the description of the present invention, "several" means one or more, "many" means more than two, "greater than," "less than," and "exceed" are understood to exclude the number itself, while "above," "below," and "within" are understood to include the number itself. The use of "first" and "second" in the description is solely for the purpose of distinguishing technical features and should not be construed as indicating or implying relative importance, implicitly specifying the number of the indicated technical features, or implicitly specifying the order of the indicated technical features.

[0027] The following describes the embodiments of the present disclosure through specific examples, and those skilled in the art can easily understand the advantages and effects of the present disclosure from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all of the embodiments. The present disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present disclosure. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present disclosure.

[0028] The purpose of the present invention is to propose a method for classifying and automatically screening big data severe pneumonia data to solve one or more technical problems existing in the prior art and at least provide a beneficial option or create conditions.

[0029] A big data classification and automated screening method for severe pneumonia uses standardized processes to collect and process data across different dimensionalities. Data collection covers patient baseline information, laboratory test data, imaging data, interventions, and clinical outcomes. Laboratory test data will be uniformly converted to international standard units, and imaging data will be stored in DICOM format to ensure compatibility across hospital systems. Metadata will be appended to all data collection processes.

[0030] In terms of data processing, all multimodal data will be cleaned and standardized to ensure data consistency, completeness and accuracy. All patients' clinical data will be recorded using standardized forms, including scores on the day of admission, including PSI, APACHE II, CURB-65 and SOFA scores, as well as dynamic changes in the disease. The data cleaning process includes checking missing values, removing duplicate data and processing outliers to ensure the quality of the data finally included in the database is reliable.

[0031] The database architecture will adopt a hybrid model, combining the strengths of relational and NoSQL databases. For structured data, relational databases such as MySQL or PostgreSQL will be used to ensure consistent and scalable data storage. For unstructured data, NoSQL databases such as MongoDB will be used to support parallel processing and distributed storage of massive amounts of data, enhancing flexibility in data expansion and retrieval.

[0032] The database will utilize a distributed storage structure to ensure data security and efficient management, supporting simultaneous access by multiple users and hospitals. The system provides a RESTful API interface to facilitate data access and sharing with external systems, enabling flexible data invocation and integration. Regarding data permission management, the system will set different access levels based on user roles to ensure the privacy of sensitive data in compliance with international privacy regulations. Patients' personal privacy information will be anonymized, and encryption technology will be used to ensure data security during storage and transmission.

[0033] Furthermore, based on the two core diagnostic data of microbial detection and host information, multi-layer simulation diagnosis is performed and a training model is output, including the following steps:

[0034] Establish the weights of two core diagnostic data: microbial detection and host information;

[0035] By comparing the weights of the file contributions in the forward and reverse weight directions, the neural network input is compared with the weights of the file contributions. The network output is determined by the forward and reverse outputs. The calculation formula is as follows:

[0036]

[0037] O t =g(Vh t +b3);

[0038] O t =g(Vh t +b3);

[0039] Wherein, x represents the input value, which is the two core diagnostic data of microbial detection and host information. The weight values ​​of the two core diagnostic data of microbial detection and host information are defined based on the multimodal fusion model. are the forward weight and reverse weight directions respectively, is the network forward output value, W1, U1, b1 are the input layer weight, hidden layer weight and bias vector of the forward output respectively, is the reverse output value of the network, W2, U2, and b2 are the input layer weight, hidden layer weight, and bias vector of the reverse output respectively. The input layer weight is the result obtained by comparing the forward output and the reverse output. L1 is the input layer weight, L2 is the hidden layer weight, and the bias vector is and If the comparison result is L2, the input layer weight, L1, the hidden layer weight, and the bias vector is and of and;

[0040] O t is the final output value, V and b3 are the output layer weight matrix and bias vector respectively, and the symbol is the splicing operation;

[0041] The final output value O t The neural network model is put into training. After the training is completed, the input data is quickly output as an indicator X. The indicator X is used to determine whether the comparison data meets the qualified standard. Each indicator X needs to be trained through a separate neural network model to determine the mapping relationship between the input data and the indicator. The files are automatically classified through the output of the neural network model and the results of the training output are output multiple times.

[0042] Furthermore, in the neural network model, the number of neurons in the network input layer is equal to the number of input data types, the number of neurons in the output layer is equal to the number of output data types, and the number of intermediate layers and the number of neurons in each layer are freely set. The setting of the intermediate layers will directly affect the accuracy of the approximate data of the neural network. Whether the number of intermediate layers or the number of neurons needs to be increased or decreased is determined based on the training results. The calculation method for determining the number of intermediate layers and the number of neurons is:

[0043] The middle layers in the neural network model are added layer by layer, and the output index X corresponding to the number of middle layers is named. When the number of middle layers is 1, the output index X1 is output, and so on. Based on the index X, a nonlinear regression function is constructed. a is a constant value, and the variance D of the function's independent variable is calculated. Where L is the number of intermediate layers at output, X i is the output index X, the X i ` is the input data corresponding to the output index X, when When A takes the minimum value, To obtain the partial derivative of the function, A is a parameter that determines the number of intermediate layers. The value of A is compared with the variance D. If D>A, the output index X is input into the neural network model, and the number of intermediate layers is increased by 1. If D≤A, the output index X is output.

[0044] Update the weight coefficients of neurons using the gradient descent method:

[0045]

[0046] in, To obtain the partial derivative of the function, P and Q are the system output error and the neuron weight increment, respectively, both of which are constant values, ω(l) is the neuron weight coefficient, Δω(l) is the updated neuron weight coefficient, δ represents the neuron learning rate, and the number of neurons set in the middle layer is determined by the Δω(l). According to the superposition analysis algorithm combined with the output value O t The algorithm is deeply learned to output top-level comparative analysis data and second-level comparative analysis data to ensure the quality of the comparative data.

[0047] Further, automatic classification is performed based on the weight of the data quality and severe pneumonia data, a memory coupled to the one or more processors, the memory storing instructions that, when executed by the one or more processors, cause the one or more processors to: provide a classification pipeline comprising a component that obtains metadata associated with a data item and existing classification metadata associated with the data item, wherein the existing classification metadata comprises a current classification value of the data item, and wherein the current classification value of the data item is stored in the data item;

[0048] A plurality of classifier modules are provided, wherein each of the plurality of classifier modules has an associated classification rule, and wherein each of the classification rules, when invoked, uses metadata associated with a data item and existing classification metadata associated with the data item to determine classification metadata for the data item, and wherein the plurality of classifier modules are mediated using at least one of: aggregate classification, authoritative classification, and high classification.

[0049] Furthermore, the method is based on a system, which includes: a processor and a memory. The processor and memory in the overall system can both run a computer program in the processor. When the processor executes the computer program, it can implement the steps of a big data severe pneumonia data classification and automatic screening method as described in any one of claims 1-4.

[0050] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete component gate circuits or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the multi-layered and hierarchical precise diagnosis system for severe pneumonia pathogens, and utilizes various interfaces and lines to connect the various sub-regions of the multi-layered and hierarchical precise diagnosis system for severe pneumonia pathogens.

[0051] The memory can be used to store the computer program and / or module, and the processor realizes the various functions of the multi-layered hierarchical severe pneumonia pathogen accurate diagnosis system by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory can mainly include a storage program area and a storage data area, wherein the storage program area can store an operating system, at least one application required for a function (such as a sound playback function, an image playback function, etc.); the storage data area can store data (such as audio data, a phone book, etc.) created according to the use of a mobile phone. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, an internal memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0052] Although the description of the present disclosure has been quite detailed and specifically describes several embodiments, it is not intended to be limited to any of these details or embodiments or any particular embodiment, so as to effectively cover the intended scope of the present disclosure. In addition, the above description of the present disclosure is based on the embodiments foreseen by the inventors, which is intended to provide a useful description, and those non-substantial changes to the present disclosure that have not yet been foreseen may still represent equivalent changes to the present disclosure.

Claims

1. A method for classifying and automatically screening large data sets of severe pneumonia, characterized in that: Standardized processes are used to collect and process data across different data types. Data collection covers patient baseline information, laboratory test data, imaging data, interventions, and clinical outcomes. Laboratory test data will be uniformly converted to international standard units, and imaging data will be stored in DICOM format to ensure compatibility across hospital systems. Metadata will be appended to all data collection processes. In terms of data processing, all multimodal data will be cleaned and standardized to ensure data consistency, completeness and accuracy. All patients' clinical data will be recorded using standardized forms, including scores on the day of admission, including PSI, APACHE II, CURB-65 and SOFA scores, as well as dynamic changes in the disease. The data cleaning process includes checking missing values, removing duplicate data and processing outliers to ensure the quality of the data finally included in the database is reliable. The database architecture will adopt a hybrid model, combining the strengths of relational and NoSQL databases. For structured data, relational databases such as MySQL or PostgreSQL will be used to ensure consistent and scalable data storage. For unstructured data, NoSQL databases such as MongoDB will be used to support parallel processing and distributed storage of massive amounts of data, enhancing flexibility in data expansion and retrieval. The database will utilize a distributed storage structure to ensure data security and efficient management, supporting simultaneous access by multiple users and hospitals. The system provides a RESTful API interface to facilitate data access and sharing with external systems, enabling flexible data invocation and integration. Regarding data permission management, the system will set different access levels based on user roles to ensure the privacy of sensitive data in compliance with international privacy regulations. Patients' personal privacy information will be anonymized, and encryption technology will be used to ensure data security during storage and transmission.

2. A method for classifying and automatically screening large amounts of severe pneumonia data according to claim 1, characterized in that: Based on the two core diagnostic data of microbial detection and host information, multi-layer simulation diagnosis is performed and the training model is output, including the following steps: Establish the weights of two core diagnostic data: microbial detection and host information; By comparing the weights of the file contributions in the forward and reverse weight directions, the neural network input is compared with the weights of the file contributions. The network output is determined by the forward and reverse outputs. The calculation formula is as follows: O t =g(Vh t +b3); O t =g(Vh t +b3); Wherein, x represents the input value, which is the two core diagnostic data of microbial detection and host information. The weight values ​​of the two core diagnostic data of microbial detection and host information are defined based on the multimodal fusion model. are the forward weight and reverse weight directions respectively, is the network forward output value, W1, U1, b1 are the input layer weight, hidden layer weight and bias vector of the forward output respectively, is the reverse output value of the network, W2, U2, and b2 are the input layer weight, hidden layer weight, and bias vector of the reverse output respectively. The input layer weight is the result obtained by comparing the forward output and the reverse output. L1 is the input layer weight, L2 is the hidden layer weight, and the bias vector is and If the comparison result is L2, the input layer weight, L1, the hidden layer weight, the bias vector is and of and; O t is the final output value, V and b3 are the output layer weight matrix and bias vector respectively, and the symbol is the splicing operation; The final output value O t The neural network model is put into training. After the training is completed, the input data is quickly output as an indicator X. The indicator X is used to determine whether the comparison data meets the qualified standard. Each indicator X needs to be trained through a separate neural network model to determine the mapping relationship between the input data and the indicator. The files are automatically classified through the output of the neural network model and the results of the training output are output multiple times.

3. A method for classifying and automatically screening large amounts of severe pneumonia data according to claim 1, characterized in that: in, In the neural network model, the number of neurons in the network input layer is equal to the number of input data types, and the number of neurons in the output layer is equal to the number of output data types. The number of intermediate layers and the number of neurons in each layer are freely set. The setting of the intermediate layers will directly affect the accuracy of the approximate data of the neural network. The need to increase or decrease the number of intermediate layers or neurons is determined based on the training results. The calculation method for determining the number of intermediate layers and neurons is as follows: The middle layers in the neural network model are added layer by layer, and the output index X corresponding to the number of middle layers is named. When the number of middle layers is 1, the output index X1 is output, and so on. Based on the index X, a nonlinear regression function is constructed. a is a constant value, and the variance D of the function's independent variable is calculated. Where L is the number of intermediate layers at output, X i is the output index X, the X i ` is the input data corresponding to the output index X, when When A takes the minimum value, To obtain the partial derivative of the function, A is a parameter that determines the number of intermediate layers. The value of A is compared with the variance D. If D>A, the output index X is input into the neural network model, and the number of intermediate layers is increased by 1. If D≤A, the output index X is output. Update the weight coefficients of neurons using the gradient descent method: in, To obtain the partial derivative of the function, P and Q are the system output error and the neuron weight increment, respectively, both of which are constant values, ω(l) is the neuron weight coefficient, Δω(l) is the updated neuron weight coefficient, δ represents the neuron learning rate, and the number of neurons set in the middle layer is determined by the Δω(l). According to the superposition analysis algorithm combined with the output value O t The algorithm is deeply learned to output top-level comparative analysis data and second-level comparative analysis data to ensure the quality of the comparative data.

4. A method for classifying and automatically screening large amounts of severe pneumonia data according to claim 1, characterized in that: Automatically classify according to the weight of the data quality and severe pneumonia data, a memory coupled to the one or more processors, the memory storing instructions that, when executed by the one or more processors, cause the one or more processors to: provide a classification pipeline comprising a component that obtains metadata associated with a data item and existing classification metadata associated with the data item, wherein the existing classification metadata comprises a current classification value of the data item, and wherein the current classification value of the data item is stored in the data item; A plurality of classifier modules are provided, wherein each of the plurality of classifier modules has an associated classification rule, and wherein each of the classification rules, when invoked, uses metadata associated with a data item and existing classification metadata associated with the data item to determine classification metadata for the data item, and wherein the plurality of classifier modules are mediated using at least one of: aggregate classification, authoritative classification, and high classification.

5. A method for classifying and automatically screening large amounts of severe pneumonia data according to claim 1, characterized in that: The method is based on a system, which includes: a processor and a memory. The processor and memory in the overall system can both run a computer program in the processor. When the processor executes the computer program, it can implement the steps of a big data severe pneumonia data classification and automatic screening method as described in any one of claims 1-4.

Images