Method and device for constructing pancreatitis subtype classification model and storage medium

Abstract

The invention relates to a pancreatitis subtype classification model construction method and device and a storage medium, and is applied to the technical field of pancreatitis subtype classification, and the method comprises the steps: obtaining clinical index data of a patient, carrying out the screening and preprocessing, and constructing a typing Gaussian mixture model, according to the method, individual time features of each patient are fitted through polynomial regression, clustering processing is carried out on parameters of the polynomial through an EM algorithm, a typing label of each patient is obtained, finally accuracy verification is carried out on output of the model, and the final pancreatitis subtype classification model is obtained. Compared with the prior art that data of genomics, metabonomics and the like needs to be analyzed in a special laboratory, the cost is lower, and due to the fact that multiple indexes at multiple time points are collected and analyzed, multidimensional and dynamic properties are added in model construction, and the accuracy of model prediction is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of pancreatitis subtype classification, and in particular to a method, device and storage medium for constructing a pancreatitis subtype classification model. Background Art

[0002] Acute pancreatitis is a common disease worldwide. About one-fifth of patients progress to severe acute pancreatitis, with a mortality rate of 20%-40%. The manifestations of acute pancreatitis are diverse, and the clinical manifestations are closely related to the treatment methods. The same treatment method has different effects on patients with different clinical manifestations. The classification of acute pancreatitis subtypes is of great significance for determining the treatment method.

[0003] Existing pancreatitis subtype classification models usually use the results of genomics and metabolomics analysis, but data such as genomics and metabolomics need to be analyzed in specialized laboratories, which are costly to use and cannot be associated with specific treatment methods. This greatly increases the difficulty of model construction and limits the practicality of the model. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a method, device and storage medium for constructing a pancreatitis subtype classification model, so as to solve the problem in the prior art that data such as genomics and metabolomics need to be analyzed in a specialized laboratory, the cost of use is high, and it cannot be associated with a specific treatment method, which greatly increases the difficulty of model construction and limits the practicability of the model.

[0005] According to a first aspect of an embodiment of the present invention, a method for constructing a pancreatitis subtype classification model is provided, the method comprising:

[0006] Acquire clinical indicator data of multiple patients, screen and preprocess the clinical indicator data to obtain standard clinical indicator data;

[0007] Constructing a classification mixed Gaussian model, inputting the standard clinical indicator data of each patient into the classification mixed Gaussian model respectively, using polynomial regression to fit the individual time characteristics of each patient, and clustering the parameters of the polynomial using the EM algorithm to obtain the classification label of each patient;

[0008] The treatment data of each patient is obtained, and the accuracy of the classification label of each patient output by the classification mixed Gaussian model is verified according to the treatment data and standard clinical indicator data of each patient. If the verification is passed, a pancreatitis subtype classification model is obtained.

[0009] Preferably,

[0010] The accuracy of the classification label of each patient output by the classification mixed Gaussian model is verified according to the treatment data and standard clinical indicator data of each patient. If the verification is passed, the pancreatitis subtype classification model is obtained, including:

[0011] The treatment data, standard clinical indicator data and classification label of each patient are input into the pre-constructed random forest classification model. If the classification result of each patient output by the pre-constructed random forest classification model is inconsistent with the classification label, the parameters of the classification mixed Gaussian model are adjusted until the classification result of each patient output by the random forest classification model is consistent with the classification label, thereby obtaining a pancreatitis subtype classification model.

[0012] Preferably, it also includes:

[0013] deploying the pancreatitis subtype classification model;

[0014] Acquiring standard clinical indicator data of the patient, and inputting the standard clinical indicator data of the patient into the pancreatitis subtype classification model;

[0015] The pancreatitis subtype classification model outputs the patient's classification results and matches the patient's treatment measures according to the patient's classification results.

[0016] Preferably, it also includes:

[0017] The standard clinical indicator data and classification results of the patients input into the pancreatitis subtype classification model are summarized according to a preset time period, and the pancreatitis subtype classification model is iteratively updated based on the summarized data.

[0018] Preferably,

[0019] The screening of the clinical indicator data comprises:

[0020] The clinical indicator data are screened to remove data whose time series is lower than a preset time series threshold and data whose number of missing values ​​is greater than a preset missing value number threshold.

[0021] Preferably,

[0022] The preprocessing of the clinical indicator data comprises:

[0023] The missing values ​​of the screened clinical indicator data were filled to obtain standard clinical indicator data.

[0024] According to a second aspect of an embodiment of the present invention, there is provided a device for constructing a pancreatitis subtype classification model, the device comprising:

[0025] Data acquisition module: used to acquire clinical indicator data of multiple patients, screen and preprocess the clinical indicator data, and obtain standard clinical indicator data;

[0026] Typing label acquisition module: used to construct a typing mixed Gaussian model, input the standard clinical indicator data of each patient into the typing mixed Gaussian model, use polynomial regression to fit the individual time characteristics of each patient, and use the EM algorithm to cluster the parameters of the polynomial to obtain the typing label of each patient;

[0027] Verification module: used to obtain the treatment data of each patient, and verify the accuracy of the classification label of each patient output by the classification mixture Gaussian model according to the treatment data and standard clinical indicator data of each patient. If the verification is passed, a pancreatitis subtype classification model is obtained.

[0028] According to a third aspect of an embodiment of the present invention, there is provided a storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a main controller, each step in the above method is implemented.

[0029] The technical solution provided by the embodiments of the present invention may have the following beneficial effects:

[0030] The present application obtains the clinical indicator data of the patients and performs screening and preprocessing, constructs a typing mixed Gaussian model, uses polynomial regression to fit the individual time characteristics of each patient, uses the EM algorithm to cluster the parameters of the polynomial, obtains the typing label of each patient, and finally verifies the accuracy of the model output to obtain the final pancreatitis subtype classification model. The model of the present application is established using common clinical routine vital signs data, and data collection is very convenient. Compared with existing genomics and metabolomics data that need to be analyzed in specialized laboratories, the cost is lower and data acquisition is more convenient. In addition, since the present application needs to fit individual time characteristics, it is necessary to collect and analyze multiple indicators at multiple time points. Multidimensionality and dynamics are added to the model construction, which improves the accuracy of model prediction.

[0031] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0033] Figure 1 is a flow chart of a method for constructing a pancreatitis subtype classification model according to an exemplary embodiment;

[0034] Figure 2 is a schematic diagram of a process architecture according to another exemplary embodiment;

[0035] Figure 3 is a system schematic diagram of a device for constructing a pancreatitis subtype classification model according to another exemplary embodiment;

[0036] In the attached figure: 1-data acquisition module, 2-typing label acquisition module, 3-verification module. DETAILED DESCRIPTION

[0037] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0038] Embodiment 1

[0039] Figure 1 is a flow chart of a method for constructing a pancreatitis subtype classification model according to an exemplary embodiment. Figure 1 As shown, the method includes:

[0040] S1, obtaining clinical indicator data of multiple patients, screening and preprocessing the clinical indicator data to obtain standard clinical indicator data;

[0041] S2, constructing a typing mixed Gaussian model, inputting the standard clinical indicator data of each patient into the typing mixed Gaussian model respectively, using polynomial regression to fit the individual time characteristics of each patient, and clustering the parameters of the polynomial using the EM algorithm to obtain the typing label of each patient;

[0042] S3, obtaining treatment data of each patient, and verifying the accuracy of the classification label of each patient output by the classification mixed Gaussian model according to the treatment data and standard clinical indicator data of each patient. If the verification is passed, a pancreatitis subtype classification model is obtained;

[0043] It can be understood that the specific content of this embodiment is:

[0044] After integrating data from multiple databases, the clinical indicator data are screened to remove data that do not meet the standards, such as removing data with a time series lower than a preset time series threshold and removing data with a number of missing values ​​greater than a preset missing value threshold, and the remaining data are preprocessed in the next step;

[0045] In data preprocessing, missing data are first filled, and the filled data are standardized;

[0046] Construct a classification mixed Gaussian model, use polynomial regression to fit individual time characteristics, and then use the EM algorithm to cluster the polynomial parameters, assign each group of data to the type that can produce the lowest MSE (minimum mean square error), and obtain the classification label. Specifically:

[0047] Assume that the data set is {(Y i,j ,X i，j )}, where i is the i-th patient, j is the j-th observation of the i-th patient, and Y i，j =(SBP i，j , DBP i,j ,HR i，j , R.R. i,j ,TEMP i，j ) is a multidimensional response variable for each individual, wherein SBP refers to the systolic blood pressure data of the patient, DBP refers to the diastolic blood pressure data of the patient, HR refers to the heart rate data of the patient, RR refers to the respiratory rate data of the patient, and TEMP refers to the body temperature data of the patient, all of which are clinical declaration sign data of the patient, derived from the above-mentioned standard clinical indicator data of the patient; is the observation time Hours and its square; then the corresponding model of K mixed components is:

[0048]

[0049] In the formula, Z i represents the mixed component category of the i-th individual, N(μ,σ 2 ) means the mean is μ and the variance is σ 2 The normal distribution, β k,SBP , β k,DBP , β k,HR , β k,SBP , β k,RR , β k,TEMP Respectively represent the coefficient vector of each response variable in the kth mixed component. By using the EM algorithm to cluster the coefficient vector, we can find which mixed component category the ith individual belongs to has the smallest error, that is, which mixed component category is the typing label of the patient.

[0050] In the verification phase, treatment data is added, and treatment data and clinical indicator data are used as input. The prediction result is the classification label assigned in the previous phase. Random forest is used as the classification model. If the classification result is the classification label obtained in the previous step, it proves that the classification in the previous phase is reasonable. If the classification result is not the classification label of the previous phase, the parameters of the classification phase are adjusted and the classification is repeated until the classification result is consistent with the classification label, thus obtaining a pancreatitis subtype classification model.

[0051] It is worth noting that, as shown in the attached Figure 2 As shown in the figure, after obtaining the pancreatitis subtype classification model, the model is deployed on the internal server and put into actual use. In the actual use process, according to the preset time period, generally 3-6 months, the standard clinical indicator data and typing results of the patients input into the pancreatitis subtype classification model are obtained. After the data of this time period are summarized, the model is further trained and verified, and the model is continuously updated to ensure the accuracy and practicality of the model. In short, the model is updated and iterated every 3-6 months;

[0052] It is worth noting that in the actual use of the model, the real-time data of patients is screened, and the data that does not meet the conditions is excluded after giving reasons. The real-time data of patients who meet the conditions is preprocessed. The missing value filling and data standardization here are based on the distribution of the training data set, and there is no need to recalculate the parameters. The processed data is transmitted to the pancreatitis subtype classification model through the intranet. The model gives the classification results and matches the corresponding treatment measures.

[0053] This embodiment uses routine vital sign data commonly used in clinical practice, and data collection is very convenient. It collects and analyzes multiple indicators at multiple time points, adds multidimensionality and dynamics to model construction, and adds treatment methods as associated states, which reduces the difficulty of model construction while improving the accuracy and practicality of the model.

[0054] Embodiment 2

[0055] Figure 3 is a system schematic diagram of a device for constructing a pancreatitis subtype classification model according to another exemplary embodiment, the device comprising:

[0056] Data acquisition module 1: used to acquire clinical indicator data of multiple patients, screen and preprocess the clinical indicator data, and obtain standard clinical indicator data;

[0057] Typing label acquisition module 2: used to construct a typing mixed Gaussian model, input the standard clinical indicator data of each patient into the typing mixed Gaussian model, use polynomial regression to fit the individual time characteristics of each patient, and use the EM algorithm to cluster the parameters of the polynomial to obtain the typing label of each patient;

[0058] Verification module 3: used to obtain the treatment data of each patient, and verify the accuracy of the classification label of each patient output by the classification mixed Gaussian model according to the treatment data and standard clinical indicator data of each patient. If the verification is passed, a pancreatitis subtype classification model is obtained.

[0059] Embodiment three:

[0060] This embodiment provides a storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a main controller, each step in the above method is implemented;

[0061] It is understandable that the storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0062] It can be understood that the same or similar parts of the above embodiments can be referenced to each other, and the contents not described in detail in some embodiments can refer to the same or similar contents in other embodiments.

[0063] It should be noted that, in the description of the present invention, the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of the present invention, unless otherwise specified, the meaning of "plurality" refers to at least two.

[0064] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention belong.

[0065] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0066] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0067] In addition, each functional unit in each embodiment of the present invention may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0068] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0069] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0070] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A method for constructing a pancreatitis subtype classification model, characterized in that: The method comprises: Acquiring clinical indicator data of multiple patients, screening and preprocessing the clinical indicator data, and obtaining standard clinical indicator data; Constructing a classification mixed Gaussian model, inputting the standard clinical indicator data of each patient into the classification mixed Gaussian model respectively, using polynomial regression to fit the individual time characteristics of each patient, and clustering the parameters of the polynomial using the EM algorithm to obtain the classification label of each patient; The treatment data of each patient is obtained, and the accuracy of the classification label of each patient output by the classification mixed Gaussian model is verified according to the treatment data and standard clinical indicator data of each patient. If the verification is passed, a pancreatitis subtype classification model is obtained.

2. The method according to claim 1, characterized in that The accuracy of the classification label of each patient output by the classification mixed Gaussian model is verified according to the treatment data and standard clinical indicator data of each patient. If the verification is passed, the pancreatitis subtype classification model is obtained, including: The treatment data, standard clinical indicator data and classification label of each patient are input into the pre-constructed random forest classification model. If the classification result of each patient output by the pre-constructed random forest classification model is inconsistent with the classification label, the parameters of the classification mixed Gaussian model are adjusted until the classification result of each patient output by the random forest classification model is consistent with the classification label, thereby obtaining a pancreatitis subtype classification model.

3. The method according to claim 2, characterized in that Also includes: deploying the pancreatitis subtype classification model; Acquiring standard clinical indicator data of the patient, and inputting the standard clinical indicator data of the patient into the pancreatitis subtype classification model; The pancreatitis subtype classification model outputs the patient's classification results and matches the patient's treatment measures according to the patient's classification results.

4. The method according to claim 3, characterized in that Also includes: The standard clinical indicator data and classification results of the patients input into the pancreatitis subtype classification model are summarized according to a preset time period, and the pancreatitis subtype classification model is iteratively updated based on the summarized data.

5. The method according to claim 4, characterized in that The screening of the clinical indicator data comprises: The clinical indicator data are screened to remove data whose time series is lower than a preset time series threshold and data whose number of missing values ​​is greater than a preset missing value number threshold.

6. The method according to claim 5, characterized in that The preprocessing of the clinical indicator data comprises: The missing values ​​of the screened clinical indicator data were filled to obtain standard clinical indicator data.

7. A device for constructing a pancreatitis subtype classification model, characterized in that: The device comprises: Data acquisition module: used to acquire clinical indicator data of multiple patients, screen and preprocess the clinical indicator data, and obtain standard clinical indicator data; Typing label acquisition module: used to construct a typing mixed Gaussian model, input the standard clinical indicator data of each patient into the typing mixed Gaussian model, use polynomial regression to fit the individual time characteristics of each patient, and use the EM algorithm to cluster the parameters of the polynomial to obtain the typing label of each patient; Verification module: used to obtain the treatment data of each patient, and verify the accuracy of the classification label of each patient output by the classification mixture Gaussian model according to the treatment data and standard clinical indicator data of each patient. If the verification is passed, a pancreatitis subtype classification model is obtained.

8. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by the main controller, each step in the method for constructing a pancreatitis subtype classification model as described in any one of claims 1 to 6 is implemented.

Images