Liver chronic disease screening method and device and storage medium

Abstract

The invention relates to the technical field of medical treatment, and discloses a liver chronic disease screening method and device and a storage medium. The method comprises the steps that a liver detection system collects liver hardness data and liver fat attenuation parameters of a target object; the mobile terminal system receives the liver hardness data and the liver fat attenuation parameters, combines the liver hardness data and the liver fat attenuation parameters with basic information of a preset target object, and generates collection feature data; the cloud service system performs label labeling matrix processing on the preprocessed integrated data to generate a label labeling matrix; performing fitting residual processing on the label marking matrix to obtain extended decision tree data; performing discrete sampling processing on the extended decision tree data to obtain a sparse vector; performing classification calculation processing on the sparse vector to obtain a classification result; and generating a disease screening report based on a classification result. In the embodiment of the invention, through data processing of the system architecture and the cloud service system, the early warning timeliness and accuracy of liver disease and chronic disease screening are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical treatment, and in particular to a liver chronic disease screening method, device and storage medium. BACKGROUND

[0002] At present, liver disease screening mainly relies on a single FibroScan device, and liver fibrosis and fatty liver degree are evaluated by measuring liver tissue hardness value (LSM) and liver fat attenuation parameter (CAP) through non-invasive liver elastography (FibroScan).

[0003] FibroScan devices are concentrated in large hospitals, and the popularization rate in primary medical institutions is low, and chronic disease screening mainly relies on clinical physical examination. However, although FibroScan detection results are mostly in the form of reports, the existing screening does not integrate a prediction model based on historical data, it is difficult to make risk classification and risk prediction on the progress of liver disease, resulting in poor timeliness of intervention. Moreover, traditional screening often makes judgments only according to single LSM or CAP value, and the analysis result is not accurate due to the dependence of single index. Therefore, the existing technology has the technical problems of insufficient intervention timeliness and inaccurate analysis result in liver chronic disease screening, and a technology is needed to solve the current problems. SUMMARY

[0004] The main purpose of the present application is to solve the technical problems of insufficient early warning timeliness and inaccurate analysis result in liver chronic disease screening.

[0005] The first aspect of the present application provides a liver chronic disease screening method, which is applied in a liver chronic disease screening system, the liver chronic disease screening system comprises a liver detection system, a mobile terminal system and a cloud service system, and the chronic disease screening method comprises the following steps: The liver detection system collects liver hardness data and liver fat attenuation parameters of a target object, and sends the liver hardness data and the liver fat attenuation parameters to the mobile terminal system; The mobile terminal system receives the liver hardness data and the liver fat attenuation parameters, combines the liver hardness data, the liver fat attenuation parameters and basic information of a preset target object, generates acquisition feature data, and displays the acquisition feature data; When a confirmation instruction is listened to, the acquisition feature data is sent to the cloud service system; The cloud service system receives the acquisition feature data, and pre-processes the acquisition feature data to obtain pre-processing integrated data; According to a preset label set, the pre-processing integrated data is label marked and matrix processed to generate a label marked matrix; According to a preset gradient boosting algorithm, the label mark matrix is subjected to fitting residual processing to obtain extended decision tree data; According to a preset sampling algorithm, the extended decision tree data is subjected to discrete sampling processing to obtain a sparse vector; According to a preset logistic regression algorithm, the sparse vector is subjected to classification calculation processing to obtain a classification result; Based on the classification result, a disease screening report is generated, and the disease screening report is sent to the mobile terminal system.

[0006] Optionally, in the first implementation manner of the first aspect of the present application, the step of generating a label mark matrix according to a preset label set and performing label mark matrix processing on the preprocessed integrated data comprises: The preprocessed integrated data is subjected to element matrix processing to obtain an integrated data matrix; Based on rows or columns, the integrated data matrix is subjected to range matching marking with a preset label set to obtain label data of rows or columns; The label data of rows or columns is combined with the integrated data matrix to generate a label mark matrix.

[0007] Optionally, in the second implementation manner of the first aspect of the present application, the step of obtaining extended decision tree data by performing fitting residual processing on the label mark matrix according to a preset gradient boosting algorithm comprises: The rows or columns corresponding to the labels of the label mark matrix are subjected to global scaling to obtain a global label matrix; Based on the label elements corresponding to the global label matrix, the rows or columns of the global label matrix are subjected to probability residual calculation to obtain a probability residual; Based on a preset CART regression tree, the rows or columns of the global label matrix are subjected to splitting processing to obtain extended decision tree data.

[0008] Optionally, in the third implementation manner of the first aspect of the present application, the step of obtaining a global label matrix by performing global scaling on the rows or columns corresponding to the labels of the label mark matrix comprises: F0(x) = log(p / (1-p)), Wherein, F0(x) is the global scaling of a single label, and p is the positive sample probability scaling of the row or column label.

[0009] Optionally, in the fourth implementation manner of the first aspect of the present application, the step of obtaining a sparse vector by performing discrete sampling processing on the extended decision tree data according to a preset sampling algorithm comprises: Based on a preset OneHotEncoder() function, the extended decision tree data is subjected to discrete feature extraction to obtain a sparse vector.

[0010] Optionally, in a fifth implementation form of the first aspect of the present application, the collection characteristic data comprises: liver stiffness, liver fat attenuation parameter, BMI parameter, gender parameter, and the step of preprocessing the collection characteristic data to obtain the preprocessed integrated data comprises: sorting the liver stiffness and the liver fat attenuation parameter in ascending order to obtain sorted liver stiffness and liver fat sorting parameter; calculating the difference between the 75% quantile data and the 25% quantile data of the sorted liver stiffness and the liver fat sorting parameter to obtain liver stiffness analysis value and liver fat analysis value; performing upper and lower limit numerical truncation processing on the sorted liver stiffness based on the liver stiffness analysis value to obtain corrected liver stiffness; performing upper and lower limit numerical truncation processing on the liver fat sorting parameter based on the liver fat analysis value to obtain corrected liver fat attenuation parameter; performing normalization processing on the BMI parameter according to a preset Z-score algorithm to obtain corrected BMI parameter; performing mapping processing on the gender parameter according to a preset mapping rule to obtain gender mapping value; combining the corrected liver stiffness, the corrected liver fat attenuation parameter, the corrected BMI parameter, and the gender mapping value to obtain the preprocessed integrated data.

[0011] Optionally, in a sixth implementation form of the first aspect of the present application, after the step of displaying the collection characteristic data, the method further comprises: when receiving an abnormal feedback instruction, receiving a feedback input instruction to generate abnormal explanation data; sending the abnormal explanation data and the collection characteristic data to the liver detection system to adjust abnormal data in the collection characteristic data.

[0012] Optionally, in a seventh implementation form of the first aspect of the present application, after the step of generating a disease screening report based on the classification result and sending the disease screening report to the mobile terminal system, the method further comprises: sending the disease screening report to the liver detection system to receive clinical feedback data returned by the liver detection system; adjusting the label set based on the clinical feedback data to obtain a new label set.

[0013] The second aspect of the present application provides a liver chronic disease screening device, comprising: a memory and at least one processor, the memory has instructions stored therein, and the memory and the at least one processor are interconnected by a circuit; the at least one processor invokes the instructions in the memory to enable the liver chronic disease screening device to perform the above-mentioned liver chronic disease screening method.

[0014] The third aspect of the present application provides a computer-readable storage medium having instructions stored therein, which, when executed on a computer, cause the computer to perform the above-mentioned liver chronic disease screening method.

[0015] In the embodiment of the present application, by setting the liver detection system, the mobile terminal system and the cloud service system, the liver tissue hardness value and the liver fat attenuation parameter collected on the liver detection system are sent to the mobile terminal system in combination with the basic information of the electronic medical record or health record of the target object. After the user of the mobile terminal system confirms the information, the collected characteristic data is sent to the cloud service system, the collected characteristic data is analyzed and calculated in the cloud service system, the high-precision disease screening result is generated, and the disease screening result is sent back to the mobile terminal system, thereby improving the early warning timeliness and accuracy of the liver chronic disease screening, and solving the technical problems of insufficient early warning timeliness and inaccurate analysis result of the liver chronic disease. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 An embodiment of the liver chronic disease screening method in the embodiment of the present application is shown in the figure; Figure 2 An embodiment of the step 103 of the liver chronic disease screening method in the embodiment of the present application is shown in the figure; Figure 3 An embodiment of the step 105 of the liver chronic disease screening method in the embodiment of the present application is shown in the figure; Figure 4 An embodiment of the step 106 of the liver chronic disease screening method in the embodiment of the present application is shown in the figure; Figure 5 An embodiment of the liver chronic disease screening device in the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0017] The embodiment of the present application provides a liver chronic disease screening method, device and storage medium.

[0018] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it is understood that the present disclosure can be embodied in various forms and should not be construed as being limited to the embodiments set forth herein, but rather should be interpreted in the context of what is claimed. It is understood that the drawings and embodiments of the present disclosure are for exemplary purposes only and are not intended to limit the scope of protection of the present disclosure.

[0019] In the description of embodiments of the present disclosure, the term "comprising" and its conjugations should be understood to encompass the meanings of "consisting of" and "consisting essentially of". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc. can refer to different or identical objects. Other explicit and implicit definitions can also be included below.

[0020] For ease of understanding, the specific process of the embodiments of the present disclosure is described below. Please refer to Figure 1 In an embodiment of the liver chronic disease screening method in the embodiments of the present disclosure, the liver chronic disease screening method is applied in a liver chronic disease screening system, the liver chronic disease screening system comprises a liver detection system, a mobile terminal system and a cloud service system, and the chronic disease screening method comprises the following steps. 101. The liver detection system collects liver stiffness data and liver fat attenuation parameters of a target object, and sends the liver stiffness data and the liver fat attenuation parameters to the mobile terminal system. In this embodiment, the collection module of the liver detection system uses a FibroScan device to collect data. FibroScan is a non-invasive liver detection device based on transient elastography technology, which can collect liver stiffness data and liver fat attenuation parameters from a target object without puncture or surgery. Finally, the liver stiffness data and the liver fat attenuation parameters collected are sent to the mobile terminal system through a network.

[0021] 102. The mobile terminal system receives the liver stiffness data and the liver fat attenuation parameters, combines the liver stiffness data, the liver fat attenuation parameters and the basic information of a preset target object, generates collection feature data, and displays the collection feature data. In the embodiment, the liver detection system sends the collected characteristic data to the mobile terminal system. The mobile terminal system can be an APP terminal of a smart phone. The APP terminal stores basic information of the target object, such as age, BMI, drinking history, metabolic syndrome related indicators, liver disease family history, etc. or the APP terminal receives the liver stiffness data and the liver fat attenuation parameter and fills in the basic information of the user. The collected information is structured and combined to generate the collected characteristic data. The collected characteristic data is displayed for the user to confirm the relevant collected characteristic data.

[0022] 103. When the confirmation instruction is monitored, the collected characteristic data is sent to the cloud service system. In the embodiment, after the APP terminal of the mobile terminal system monitors the confirmation instruction of the collected characteristic data, the collected characteristic data is sent to the cloud service system for feature analysis.

[0023] Further, after the step 102 of displaying the collected characteristic data, the following steps are further included. 1031. When the abnormal feedback instruction is received, the feedback input instruction is received to generate abnormal description data. 1032. The abnormal description data and the collected characteristic data are sent to the liver detection system to adjust the abnormal data in the collected characteristic data.

[0024] In the steps 1031-1032, the abnormal feedback instruction of the collected characteristic data in the mobile terminal system is received. The abnormal feedback instruction can input the description of the abnormal data, such as missing liver stiffness data or liver fat attenuation parameter exceeding the normal range. Then the abnormal description data and the collected characteristic data are sent to the liver detection system. Based on the abnormal description data, the management personnel of the liver detection system adjusts the abnormal data in the collected characteristic data in combination with the basic information of the target object in the collected characteristic data to timely solve the abnormal situation of the measurement data in the collection terminal. An encryption communication protocol is used in the data transmission process to ensure the safety of the detection results and personal information.

[0025] 104. The cloud service system receives the collected characteristic data and pre-processes the collected characteristic data to obtain pre-processed integrated data. In the embodiment, the cloud service system receives the collection characteristic data, and the structure of the collection characteristic data can be a combination of liver stiffness data (LSM), liver fat attenuation parameter (LSM), BMI, age, gender, lifestyle, and the like, for example, an ordering structure combined by columns of [[liver stiffness data', 'liver fat attenuation parameter', 'age', 'gender', 'BMI', 'lifestyle']], and of course, the ordering structure combined by rows can also be formed, which is only a transposed form of the above data.

[0026] Specifically, please refer to Figure 2 , Figure 2 For a specific embodiment of the 104th step of the liver chronic disease screening method in the embodiment, the collection characteristic data includes: liver stiffness, liver fat attenuation parameter, BMI parameter, and gender parameter, and the step of “preprocessing the collection characteristic data to obtain preprocessed integrated data” in the 104th step includes the following specific implementation: 1041. The liver stiffness and the liver fat attenuation parameter are sorted in ascending order to obtain sorted liver stiffness and liver fat sorting parameters. 1042. The difference between the 75% quantile data and the 25% quantile data of the sorted liver stiffness and the liver fat sorting parameter is calculated to obtain liver stiffness analysis value and liver fat analysis value. 1043. The sorted liver stiffness is truncated based on the upper and lower limit values of the liver stiffness analysis value to obtain a corrected liver stiffness. 1044. The liver fat attenuation parameter is truncated based on the upper and lower limit values of the liver fat analysis value to obtain a corrected liver fat attenuation parameter. 1045. The BMI parameter is normalized according to a preset Z-score algorithm to obtain a corrected BMI parameter. 1046. The gender parameter is mapped according to a preset mapping rule to obtain a gender mapping value. 1047. The corrected liver stiffness, the corrected liver fat attenuation parameter, the corrected BMI parameter, and the gender mapping value are combined to obtain preprocessed integrated data.

[0027] In the 1041-1047 steps, the preprocessing method of the liver stiffness and the liver fat attenuation parameter is similar. The 1041-1044 steps are explained by taking the liver stiffness as an example, and the preprocessing of the liver fat attenuation parameter is referred to the preprocessing of the liver stiffness.

[0028] For example, the liver stiffness detection data of a batch is [4.2, 5.1, 6.3, 7.5, 8.2, 25.6, 9.1] (unit: kPa), and the sorted liver stiffness is [4.2, 5.1, 6.3, 7.5, 8.2, 9.1, 25.6] after sorting.

[0029] In the sorted liver stiffness, the data of the 25th percentile is 5.1, and the data of the 75th percentile is 9.1, and the difference between the two is 4.0, that is, the liver stiffness analysis value is 4.

[0030] The lower limit value is set to the data of the 25th percentile minus 1.5 times the liver stiffness analysis value, and the lower limit value is -0.9, which is adjusted to 0 because the liver stiffness cannot be negative.

[0031] The upper limit value is set to the data of the 75th percentile plus 1.5 times the liver stiffness analysis value, and the upper limit value is 15.1, and all data greater than 15.1 in the sorted liver stiffness are corrected to 15.1.

[0032] The final corrected liver stiffness is [4.2, 5.1, 6.3, 7.5, 8.2, 9.1, 15.1], and the abnormal value modification method of liver fat attenuation parameter is similar.

[0033] The BMI data is converted into a distribution with a mean of 0 and a standard deviation of 1 using Z-score standardization to ensure that the model is fair to the weights of each feature. The mean ( ) and standard deviation ( ) of the BMI data are calculated, and each data point is standardized.

[0034] For example, the mean of the BMI data of a certain group is =24.5, and the standard deviation is =3.2, and the BMI of a certain user is 28.3.

[0035] The standardized value is (28.3-24.5) / 3.2=1.1875, according to the mathematical formula: , where is the original value of the BMI data, is the mean of the BMI data of the group, is the standard deviation of the BMI data of the group, is the standardized value. After the above processing, the corrected BMI parameter is obtained by normalizing the BMI parameter.

[0036] Set the mapping rule to encode the male mapping value as 1 and the female mapping value as 0, and the gender parameter is mapped to obtain the corresponding gender mapping value.

[0037] The corrected liver stiffness, the corrected liver fat attenuation parameter, the corrected BMI parameter, and the gender mapping value generated above are combined according to a data structure set to generate preprocessed integrated data.

[0038] 105. According to the preset label set, the preprocessed integrated data is subjected to label marking matrix processing to generate a label marking matrix. In this embodiment, the data structure of the [[liver stiffness data', 'liver fat attenuation parameter', 'age', 'gender', 'BMI', 'lifestyle']] is sequentially subjected to range data matching analysis in the preset label set to obtain corresponding labels such as [5, 9.5, 15, 0, 0.7, 36]. Each row of sample has one-to-one correspondence with 5 values and 1 label. The label is integrated into the preprocessed integrated data to generate a label marking matrix.

[0039] Specifically, please refer to Figure 3 , Figure 3 As a specific embodiment of step 105 of the liver chronic disease screening method in the embodiment of the present application, the following specific embodiments are included in step 105: 1051. The preprocessed integrated data is subjected to element matrix processing to obtain an integrated data matrix. 1052. Based on rows or columns, the integrated data matrix is subjected to range matching marking with a preset label set to obtain label data of rows or columns. 1053. The label data of rows or columns is combined with the integrated data matrix to generate a label marking matrix.

[0040] In steps 1051-1053, the preprocessed integrated data can exist in the form of a dictionary. The dictionary elements are subjected to matrix processing to obtain an integrated data matrix in the form of a list.

[0041] According to the data structure form of the integrated data matrix, various data are sorted in the form of columns. The column data of the integrated data matrix are used to perform range matching marking with the preset label set. Various data are arranged in the form of rows. The row data of the integrated data matrix are used to perform range matching marking with the preset label set to obtain label data of row or column combinations. According to the one-to-one correspondence between the label data and the entity data in the integrated data matrix, a row or column is added in the integrated data matrix as a label to add label data, thereby generating a label marking matrix.

[0042] 106. According to a preset gradient boosting algorithm, the label marking matrix is subjected to residual fitting processing to obtain extended decision tree data. In this embodiment, the marking labels in the label marking matrix are subjected to global probability calculation to generate a global proportion of each sample label as the prediction value F0(x i). The i-th sample m-th round prediction value Fm(x m (x i ), is converted into the i-th sample m-th round estimated probability y i ^ = σ (Fm(x m (x i )), then the i-th sample true y i (take 0 or 1), and finally the m-th round residual r i = y i - y i ^ , the greater the residual r i , the lower the model estimates or overestimates. Multiple rounds of data fitting, the residual r i and sample i are divided into 3-layer regression trees using squared error, and the residual mean of each split leaf is saved as the initial residual value. Iterative loop generates decision tree, and finally obtains extended decision tree data.

[0043] Specifically, please refer to Figure 4 , Figure 4 is a specific embodiment of the 106 steps of the liver chronic disease screening method in the embodiment of the application, and the 106 steps include the following specific embodiments: 1061. Global scaling is performed on the rows or columns of the label marking matrix corresponding to the labels to obtain a global label matrix; 1062. Probability residual calculation is performed on the rows or columns of the global label matrix based on the label elements corresponding to the global label matrix to obtain a probability residual; 1063. Split processing is performed on the rows or columns of the global label matrix based on a preset CART regression tree to obtain extended decision tree data.

[0044] In the steps 1061-1063, assuming that there are three samples, label 1 is set as {1, 0, 1}, and label 2 is set as {0, 0, 1}, the global proportion of the three positive samples of label 1 is 2 / 3, and the global proportion of the three positive samples of label 2 is 1 / 3. The global proportion log(p / (1-p)) of label 1 and label 2 is calculated respectively, p is 2 / 3 and 1 / 3, 0.6931 and-0.6931 are obtained. The obtained global proportions of label 1 and label 2 replace the original labels, label 1 becomes {0.6931, 0.6931, 0.6931} and label 2 becomes {-0.6931, -0.6931, -0.6931}, and a global label matrix is obtained. Then, the global proportion of each sample label in the global label matrix is set as the 0-th round prediction value F0(x i of the i-th sample, the m-th round prediction value Fm(x m of the i-th sample is calculated, and the i-th sample m-th round prediction value Fm(x i), convert to the estimated probability y of the i-th sample in the m-th round i ^ = sigma (F m (x i )), then the real y of the i-th sample i (take 0 or 1), and finally get the residual r of the m-th round i =y i -y i ^ , the larger the residual r i , the more the model underestimates or overestimates. Multiple rounds of data fitting, based on the preset CART regression tree, the residual r i and the sample i are divided into 3-layer regression trees using squared error, and the "residual mean of the leaf" is saved in each split leaf as the initial residual value r i , and the decision tree is generated by iteration, and the residual is fitted by CART regression tree, and each tree has at most 3 layers, so there are at most 8 leaves, and an optimal step learning rate γ is learned for the samples falling into the same leaf, and F m (x i ) = F m-1 (x i ) + 0.1 * gamma * tree, tree is the sample weight equal to [r1, r2,..., r i ]^T, output the new adjusted sample weight, save 100 trees, generate 3 layers of data, each tree has 8 leaves (may be less), the total number of leaves is less than 800, and finally get the extended decision tree data.

[0045] Specifically, the step of globally scaling the row or column corresponding to the label of the label mark matrix to obtain a global label matrix comprises: F0(x) = log(p / (1-p)), Where F0(x) is the global proportion of a single label, and p is the positive sample probability proportion of the row or column label.

[0046] In this embodiment, it is assumed that there are 3 samples, label 1 is set to {1, 0, 1}, and label 2 is set to {0, 0, 1}. The global proportion of the three positive samples of label 1 is 2 / 3, and the global proportion of the three positive samples of label 2 is 1 / 3. The global proportion log(p / (1-p)) of label 1 and label 2 is calculated respectively, p is 2 / 3 and 1 / 3, and 0.6931 and -0.6931 are obtained. The global proportions of label 1 and label 2 obtained are replaced with the original labels, label 1 becomes {0.6931, 0.6931, 0.6931} and label 2 becomes {-0.6931, -0.6931, -0.6931}, and a global label matrix is obtained.

[0047] 107. Discretely sample the extended decision tree data according to a preset sampling algorithm to obtain a sparse vector; In this embodiment, for each sample 100 trees each give "fall in which leaf", 8 leaves for each tree, 100 trees at most 800 features (actual sparse, lack of leaves without corresponding column), if the test set appears the leaf number that the training set has not seen, directly map to the all-0 vector. After the decision tree is discretely sampled, a sparse vector that only stores non-zero values is obtained. It should be noted that the sampling logic and method are obtained by training.

[0048] Specifically, the following specific implementation is included in step 107: 1071. Based on the preset OneHotEncoder() function, the extended decision tree data will be discretely feature extracted to obtain a sparse vector.

[0049] In step 1071, the OneHotEncoder() function in the sci-sklearn module can be used to discretely sample the extended decision tree data. OneHotEncoder(handle_unknown='ignore') here performs discrete feature extraction and is set to a function that directly maps to an all-0 vector when the test set appears a leaf number that the training set has not seen. Based on OneHot encoding, the sparse vector is extracted to the memory for operation.

[0050] 108. According to a preset logistic regression algorithm, the sparse vector is classified and calculated to obtain a classification result; In this embodiment, the sparse vector is linearly scored and compressed by a Sigmoid function, and the samples are divided into corresponding categories according to a pre-set threshold to obtain a classification result. In the scikit-learn module, the LogisticRegression model can be used to classify and train the sparse vector, and the penalty can be set to L2 penalty, L2=1 / C, and the penalty coefficient C is set to 2.

[0051] 109. Based on the classification result, a disease screening report is generated, and the disease screening report is sent to the mobile terminal system.

[0052] In this embodiment, based on the classification result obtained by the above trained chronic disease classification analysis, the classification result, user information, and collected characteristic data are replaced into the placeholders according to the original template to generate a basic report, and a large model with disease analysis knowledge base is called to analyze the entire classification result to generate analysis text. The analysis text is added to the basic report to generate a disease screening report. Finally, the disease screening report is sent to the mobile terminal system so that the user can receive the disease screening report in time and view the screening report on the APP terminal.

[0053] Specifically, after step 109, the following specific embodiments are further included: 1091, sending the disease screening report to the liver detection system and receiving the clinical feedback data returned by the liver detection system; 1092, adjusting the label set based on the clinical feedback data to obtain a new label set.

[0054] In steps 1091-1092, the report content includes detection indicators (LSM / CAP), risk scores, classification results, and personalized suggestions (such as lifestyle adjustments, follow-up suggestions, and retest intervals). At the same time, the system pushes abnormal risk reminders to users or medical staff, sends the disease screening report to the liver detection system, and doctors or health management personnel can view the aggregated screening data through the cloud platform and return the clinical feedback data of each disease screening report through clinical observation.

[0055] The cloud service system can modify the label set matched with the collected characteristic data based on the returned clinical feedback data to obtain a new label set, so as to timely adjust the label set and improve the risk early warning accuracy, and provide a basis for actual deployment and optimization.

[0056] In the embodiment of the present application, by setting the liver detection system, the mobile terminal system, and the cloud service system, the liver tissue hardness value and the liver fat attenuation parameter are collected on the liver detection system, and the basic information of the target object electronic medical record or health record is sent to the mobile terminal system. After the user of the mobile terminal system confirms the information, the collected characteristic data is sent to the cloud service system, the characteristic analysis and regression calculation of the collected characteristic data are performed in the cloud service system, the high-precision disease screening result is generated, and the disease screening result is sent back to the mobile terminal system. The early warning timeliness and accuracy of liver disease chronic disease screening are improved, and the technical problems of insufficient early warning timeliness and inaccurate analysis result of liver disease chronic disease are solved.

[0057] Figure 5Fig. 1 is a schematic diagram of a structure of a liver chronic disease screening device according to an embodiment of the present application. The liver chronic disease screening device 500 can have a great difference in configuration or performance, and can include one or more central processing units (CPUs) 510 (e.g., one or more processors) and a memory 520, and one or more storage media 530 (e.g., one or more mass storage devices) storing applications 533 or data 532. The memory 520 and the storage media 530 can be temporary storage or persistent storage. The programs stored in the storage media 530 can include one or more modules (not shown in the figure), each of which can include a series of instruction operations in the liver chronic disease screening device 500. Further, the processor 510 can be configured to communicate with the storage media 530 to execute the series of instruction operations in the storage media 530 on the liver chronic disease screening device 500.

[0058] The liver chronic disease screening device 500 can further include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input / output interfaces 560, and / or one or more operating systems 531, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art can understand that the liver chronic disease screening device 500 can include more or fewer components than those shown, or a combination of some components, or different arrangement of components. Figure 5 The structure of the liver chronic disease screening device shown does not constitute a limitation on the liver chronic disease screening device, and can include more or fewer components than those shown, or a combination of some components, or different arrangement of components.

[0059] The present application also provides a computer readable storage medium, which can be a non-volatile computer readable storage medium or a volatile computer readable storage medium, and the computer readable storage medium stores instructions, which, when executed on a computer, cause the computer to perform the steps of the liver chronic disease screening method.

[0060] In the context of this disclosure, a machine-readable medium can be a tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include but is not limited to an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium will include one or more lines of electrical connections, portable computer disks, hard disk drives, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the foregoing.

[0061] Moreover, while operations have been depicted in a particular order, this should not be understood as requiring such order nor that all illustrated operations be performed, to achieve desirable results. In certain circumstances, multitasking and parallel processing can be advantageous. Likewise, while several specific implementation details have been discussed, these should not be construed as limitations on the scope of the disclosure. Certain features that are described in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable sub-combination.

[0062] Although the subject matter has been described in language specific to structural features and / or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims.

Claims

1. A method for screening chronic liver diseases, characterized in that: The chronic liver disease screening method is applied in a chronic liver disease screening system, which includes: a liver detection system, a mobile terminal system, and a cloud service system. The chronic disease screening method includes: The liver detection system collects liver hardness data and liver fat attenuation parameters of the target object, and sends the liver hardness data and the liver fat attenuation parameters to the mobile terminal system; The mobile terminal system receives the liver stiffness data and the liver fat attenuation parameter, combines the liver stiffness data and the liver fat attenuation parameter with basic information of a preset target object, generates collected feature data, and displays the collected feature data; When a confirmation instruction is monitored, the collected feature data is sent to the cloud service system; The cloud service system receives the collected feature data, and pre-processes the collected feature data to obtain pre-processed integrated data; According to a preset label set, the pre-processed integrated data is subjected to label matrix processing to generate a label matrix; According to a preset gradient boosting algorithm, fitting residual processing is performed on the label marking matrix to obtain extended decision tree data; Performing discrete sampling processing on the extended decision tree data according to a preset sampling algorithm to obtain a sparse vector; Performing classification calculation on the sparse vector according to a preset logistic regression algorithm to obtain a classification result; Based on the classification result, a disease screening report is generated, and the disease screening report is sent to the mobile terminal system.

2. The method for screening chronic liver diseases according to claim 1, characterized in that: The step of performing label matrix processing on the pre-processed integrated data according to the preset label set to generate a label matrix includes: Performing element matrix processing on the pre-processed integrated data to obtain an integrated data matrix; Based on the rows or columns, the integrated data matrix is ​​range-matched with the preset label set to obtain label data of the rows or columns; The label data of the row or column is merged with the integrated data matrix to generate a label marking matrix.

3. The method for screening chronic liver diseases according to claim 1, characterized in that: The step of performing fitting residual processing on the label marking matrix according to a preset gradient boosting algorithm to obtain extended decision tree data includes: Globally scaling the rows or columns of the label marking matrix corresponding to the labels to obtain a global label matrix; Based on the label elements corresponding to the global label matrix, performing probability residual calculation on the rows or columns of the global label matrix to obtain probability residuals; Based on the preset CART regression tree, the rows or columns of the global label matrix are split to obtain extended decision tree data.

4. The method for screening chronic liver diseases according to claim 3, characterized in that: The step of globally scaling the rows or columns of the label marking matrix corresponding to the labels to obtain a global label matrix comprises: F0(x)=log(p / (1-p)), Among them, F0(x) is the global proportion of a single label, and p is the probability ratio of positive samples of row or column labels.

5. The method for screening chronic liver diseases according to claim 1, characterized in that: The step of performing discrete sampling processing on the expanded decision tree data according to a preset sampling algorithm to obtain a sparse vector includes: Based on the preset OneHotEncoder() function, discrete features are extracted from the extended decision tree data to obtain a sparse vector.

6. The method for screening chronic liver diseases according to claim 1, characterized in that: The collected characteristic data includes: liver stiffness, liver fat attenuation parameter, BMI parameter, and gender parameter. The step of preprocessing the collected characteristic data to obtain preprocessed integrated data includes: Sorting the liver stiffness and the liver fat attenuation parameters from small to large to obtain ranked liver stiffness and liver fat ranking parameters; Calculating the difference between the 75% quantile data and the 25% quantile data of the ranked liver stiffness and the ranked liver fat parameters to obtain a liver stiffness analysis value and a liver fat analysis value; Based on the liver stiffness analysis value, performing upper and lower limit numerical truncation processing on the ranked liver stiffness to obtain a corrected liver stiffness; Based on the liver fat analysis value, performing upper and lower limit numerical truncation processing on the liver fat ranking parameter to obtain a modified liver fat attenuation parameter; Normalizing the BMI parameter according to a preset Z-score algorithm to obtain a corrected BMI parameter; Mapping the gender parameter according to a preset mapping rule to obtain a gender mapping value; The corrected liver stiffness, the corrected liver fat attenuation parameter, the corrected BMI parameter, and the gender mapping value are combined to obtain preprocessed integrated data.

7. The method for screening chronic liver diseases according to claim 1, characterized in that: After the step of displaying the collected feature data, the method further includes: When an abnormal feedback instruction is received, a feedback input instruction is received and abnormal description data is generated; The abnormality description data and the acquired characteristic data are sent to the liver detection system so as to adjust the abnormal data in the acquired characteristic data.

8. The method for screening chronic liver diseases according to claim 1, characterized in that: After the steps of generating a disease screening report based on the classification result and sending the disease screening report to the mobile terminal system, the method further includes: sending the disease screening report to the liver detection system and receiving clinical feedback data returned by the liver detection system; Based on the clinical feedback data, the label set is adjusted to obtain a new label set.

9. A chronic liver disease screening device, characterized in that: The chronic liver disease screening device comprises: a memory and at least one processor, wherein instructions are stored in the memory, and the memory and the at least one processor are interconnected via a line; The at least one processor calls the instructions in the memory to enable the chronic liver disease screening device to perform the chronic liver disease screening method according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for screening chronic liver diseases according to any one of claims 1 to 8 is implemented.

Images