Obstructive jaundice liver injury prediction method and system

Abstract

The invention relates to the technical field of medical data processing, in particular to an obstructive jaundice liver injury prediction method and system.The method comprises the steps that obtained prediction indexes are classified into first-class indexes, second-class indexes and third-class indexes according to specific situations of literature statistics prediction indexes; the first-class indexes are not screened, the second-class indexes are screened through a correlation analysis method, and the third-class indexes are screened through a prediction contribution degree added value and appearance time comprehensive screening method; the screening scheme is designed according to the specific situation of the prediction index, so that the index screening accuracy is improved; meanwhile, the situation of precision damage caused by improper selection of screening parameters due to adoption of a complex model screening method is also avoided; meanwhile, when the three types of indexes are screened, according to the characteristics of the three types of indexes, the three types of indexes are comprehensively screened by introducing two parameters, namely the predicted contribution degree added value and the occurrence time, so that the index screening accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data processing, and in particular to a method and system for predicting obstructive jaundice liver damage. Background Art

[0002] Obstructive jaundice refers to the obstruction of the bile duct by stones in the bile duct or tumors in the bile duct and adjacent areas, which causes poor bile flow into the duodenum and increases the pressure in the bile duct. This in turn increases the level of conjugated bilirubin in the blood, leading to symptoms such as yellowing of the skin and sclera, yellow urine, etc. It is a common disease in surgical clinics. Obstructive jaundice can cause liver damage, inflammation, decreased intestinal barrier function, endotoxemia, coagulation dysfunction, decreased immune function, and malnutrition.

[0003] There are technical solutions for predicting liver damage in the prior art. For example, a Chinese invention patent (CN112634996A) discloses a method for predicting liver damage. The method determines the weights of the chemical components of different drugs, inputs the chemical components of different drugs and the corresponding weights into a pre-trained liver damage prediction model, and obtains a prediction result of whether the combination of drugs will cause liver damage. When the drugs undergo metabolic reactions, the weights of the chemical components of the metabolites are determined based on the chemical components of the metabolites, thereby predicting whether the metabolites will cause liver damage.

[0004] However, the above scheme does not involve liver injury indicators when predicting liver injury, which affects the model prediction efficiency and accuracy. At the same time, the existing technology generally uses optimization algorithms such as Bayesian optimization model to optimize the hyperparameters of the liver injury prediction model, resulting in complex model hyperparameter design and poor operability. Summary of the Invention

[0005] In view of this, the present invention proposes a method and system for predicting obstructive jaundice liver damage, which are used to improve the efficiency and accuracy of predicting obstructive jaundice liver damage.

[0006] In order to achieve the above object, a method for predicting liver damage caused by obstructive jaundice is provided, the method comprising: S1: Obtaining predictive indicators for predicting obstructive icteric liver injury; S2: Classify the predicted indicators; The S2 is specifically: S2.1: Count the percentage of occurrences of the prediction indicators in the prediction model; S2.2: The prediction indicators are classified into first-class indicators, second-class indicators and third-class indicators according to the percentage of occurrence; S3: screening the prediction indicators according to the classification results of S2 to obtain the final indicators for predicting obstructive icteric liver injury; The S3 specifically includes: not performing a screening operation on the first type of indicators; performing a screening operation on the second type of indicators using a correlation analysis method; and performing a screening operation on the third type of indicators using a comprehensive screening method of predicted contribution increase value and appearance time; S4: establishing a prediction model for liver injury caused by obstructive jaundice; wherein the prediction model for liver injury caused by obstructive jaundice is an artificial neural network model; S5: Obtaining a training set for training the artificial neural network model; S6: training the artificial neural network model using the training set of the artificial neural network model to obtain a trained artificial neural network model; S7: Inputting the specific numerical value of the final indicator for predicting obstructive jaundice liver damage of the object to be predicted into the trained artificial neural network model to obtain a prediction result for obstructive jaundice liver damage.

[0007] Preferably, in S3, the screening operation for the two types of indicators using the correlation analysis method is specifically as follows: obtaining a data set for predicting obstructive jaundice liver damage, wherein the data set at least includes the prediction indicators in S1 and the probability of obstructive jaundice liver damage determined by experts based on the prediction indicators, and obtaining the correlation factor between the indicator to be screened and the probability of obstructive jaundice liver damage in the data set. If the correlation factor is less than 0.8, the indicator to be screened is deleted.

[0008] Preferably, in S3, the screening operation for the three types of indicators using a comprehensive screening method of predicted contribution increase value and appearance time is specifically as follows: Sa: records the year when the indicator to be screened first appears in the literature; Sb: Get the appearance time y of the indicator to be filtered i , where i is the number of the indicator to be screened; Sc: Get the predicted contribution increase value a of the index to be screened to the predicted result i ; Sd: The appearance time y of the index to be screened i And the predicted contribution value of the index to be screened to the prediction result a i The indicators to be screened are screened.

[0009] Preferably, in the calculation process of the predicted contribution increase value: all indicators in a category of indicators are input as prediction indicators into the obstructive liver injury prediction model to predict obstructive liver injury, and the prediction accuracy of the category of indicators is obtained; then, the category of indicators and the indicators to be screened are input as prediction indicators into the obstructive liver injury prediction model to predict obstructive liver injury, and the prediction accuracy of the category of indicators and the indicators to be screened is obtained; the difference a between the prediction accuracy of the category of indicators and the indicators to be screened and the prediction accuracy of the category of indicators is calculated. i As the predicted contribution increase value of the indicator i to be screened.

[0010] Preferably, the Sd is specifically: the appearance time y of the index to be screened i The predicted contribution increase value a of the to-be-screened indicator to the predicted result is less than the first threshold value i If the value is greater than the second threshold, the indicator to be screened is retained.

[0011] Preferably, the final indicators for predicting obstructive icteric liver injury after the S3 screening are procalcitonin, total bilirubin content, γ-glutamyl transpeptidase to platelet ratio, activated partial thromboplastin time, and lactic acid content.

[0012] Preferably, in S2.2, the first category of indicators are indicators with a percentage of occurrence in the interval [0.90, 1.00), the second category of indicators are indicators with a percentage of occurrence in the interval [0.75, 0.90), and the third category of indicators are indicators with a percentage of occurrence less than 0.75.

[0013] Preferably, in the Sb, the occurrence time y i It is the difference between 2004 and the year when the indicator to be screened first appears.

[0014] Preferably, the Sd is specifically: the appearance time y of the index to be screened i The predicted contribution increase value a of the to-be-screened indicator to the predicted result is less than the first threshold value i If the value is greater than the second threshold, the indicator to be screened is retained. According to another aspect of the present invention, a system for predicting liver damage caused by obstructive jaundice is provided. The system adopts the above-mentioned method for predicting liver damage caused by obstructive jaundice, and the system comprises: A prediction index acquisition module, used to acquire prediction indexes for predicting obstructive icteric liver injury; A prediction indicator classification module, used for classifying the prediction indicators; a final indicator determination module, configured to screen the prediction indicators according to the classification results of the prediction indicator classification module to obtain a final indicator for the prediction of obstructive icteric liver injury; Prediction model building module, used to establish a prediction model for obstructive jaundice liver injury; A training set acquisition module, used to acquire a training set for training the artificial neural network model; A training module, configured to train the artificial neural network model using the training set of the artificial neural network model to obtain a trained artificial neural network model; The prediction module is used to input the specific numerical value of the final indicator of the obstructive jaundice liver damage prediction of the object to be predicted into the trained artificial neural network model to obtain the obstructive jaundice liver damage prediction result.

[0015] The advantages and beneficial effects of the present invention are: The present invention classifies the obtained prediction indicators according to the specific circumstances of the literature statistical prediction indicators into first-class indicators, second-class indicators and third-class indicators; first-class indicators are not screened, second-class indicators are screened by using the correlation analysis method, and third-class indicators are screened by using the comprehensive screening method of predicted contribution increase value and appearance time; the screening scheme is designed according to the specific circumstances of the prediction indicators, thereby improving the accuracy of the indicator screening; and at the same time, it also avoids the situation where the accuracy is damaged due to improper selection of screening parameters caused by the use of complex model screening methods; At the same time, when screening the three types of indicators, the present invention introduces two parameters, namely, the predicted contribution increase value and the occurrence time, to comprehensively screen the three types of indicators according to their characteristics, thereby improving the accuracy of indicator screening. At the same time, the present invention establishes a connection between the value of the number of iterations and the learning rate and the number of prediction indicators. When the learning rate is small, a larger number of iterations is designed. When there are more prediction indicators, a larger number of iterations is also designed. That is, the number of iterations is adjusted by the learning rate and the number of prediction indicators. In this way, the model training efficiency and training accuracy can be taken into account in the model training link. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the description of the present invention or the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 A flowchart of a method for predicting liver damage caused by obstructive jaundice provided in an embodiment of the present invention; Figure 2 A flow chart for classifying the prediction indicators provided in an embodiment of the present invention; Figure 3A flowchart of a screening operation for three types of indicators using a comprehensive screening method of predicted contribution increase value and appearance time provided by an embodiment of the present invention; Figure 4 This is a structural diagram of the artificial neural network model provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0019] As attached Figure 1 As shown, a method for predicting liver damage caused by obstructive jaundice comprises the following steps: S1: Obtaining predictive indicators for predicting obstructive icteric liver injury; In this embodiment, the prediction indicators of the existing liver injury prediction model are determined based on literature search; Specifically, the keywords "obstructive jaundice" and "liver injury" were searched on literature search websites (such as CNKI and Wanfang), and the hit literature was downloaded and read, and the existing predictive indicators used in the prediction model of obstructive jaundice liver injury were recorded; In this step, the indicators include: alanine aminotransferase content (ALT), aspartate aminotransferase (AST), serum albumin content (ALB), procalcitonin (PCT), total bilirubin content (TBIL), direct bilirubin content (DBIL), unconjugated bilirubin content (IBIL), aspartate aminotransferase to alanine aminotransferase ratio (AST / ALT), gamma-glutamyl transpeptidase (γ-GT), alkaline phosphatase (ALP), uric acid content (UA), lactate dehydrogenase content (CLDH), total protein (TP ), albumin-globulin ratio (A / G), globulin content (GLB), prealbumin (PA), total bile acid (TBA), white blood cell count (WBC), red blood cell count (RBC), platelet count (PLT), lymphocyte count (LYMPH), basophil count (BASO), neutrophil count (NEUT); γ-glutamyl transpeptidase to platelet ratio (GPR), prothrombin time (PT), activated partial thromboplastin time (APTT), lactate content (LAO) and Sequential Organ Failure Assessment (SOFA); S2: Classify the predicted indicators; Specifically, as attached Figure 2 As shown, the S2 is specifically: S2.1: Count the percentage of occurrences of the prediction indicators in the prediction model; According to the hit literature in S1, relevant articles using prediction models for liver injury prediction are counted, wherein the prediction models include: fitting models, regression models, deep learning, ensemble learning, transfer learning, etc., and then the number of times the prediction indicators appear in the prediction models is counted, and finally the percentage of the number of times the prediction indicators appear in the prediction models is calculated based on the number of occurrences and the total number of prediction schemes; S2.2: The prediction indicators are classified into first-class indicators, second-class indicators and third-class indicators according to the percentage of occurrence; Among them, the first category of indicators is the indicator with a percentage of occurrence in the interval [0.90, 1.00), the second category of indicators is the indicator with a percentage of occurrence in the interval [0.75, 0.90), and the third category of indicators is the indicator with a percentage of occurrence less than 0.75; S3: screening the prediction indicators according to the classification results of S2 to obtain the final indicators for predicting obstructive icteric liver injury; Since relevant research is relatively extensive, the above steps obtain relatively more prediction indicators. For the prediction of obstructive jaundice liver injury, the prediction effectiveness of different prediction indicators is not the same. Some indicators have relatively high prediction effectiveness, while some indicators have relatively low prediction effectiveness. If all indicators are input into the prediction model, especially the deep learning related model, on the one hand, the more indicators, the more complex the training and calculation of the model, and even the complexity will increase exponentially. On the other hand, the more indicators, the less contribution some indicators make to the accuracy of the model prediction results, affecting the final prediction results of the model, and thus causing the accuracy of the final prediction results to be affected to a certain extent. Therefore, in this embodiment, the prediction accuracy and prediction efficiency of the prediction model are improved by performing a classification operation on the indicators in step S1. Specifically, the S3 is as follows: for the first type of indicators, no screening operation is performed; for the second type of indicators, a correlation analysis method is used to perform a screening operation; for the third type of indicators, a comprehensive screening method of predicted contribution increase value and appearance time is used to perform a screening operation; The correlation analysis method is a statistical analysis method for studying the correlation between two or more random variables. In this step, it is mainly used to study the correlation between the prediction index and the probability of occurrence of obstructive jaundice liver injury. The two types of indicators are screened using the correlation analysis method as follows: obtaining a data set for predicting obstructive jaundice liver injury, wherein the data set at least includes the prediction index in S1 and the probability of occurrence of obstructive jaundice liver injury determined by experts based on the prediction index, and obtaining the correlation factor between the to-be-screened index and the probability of occurrence of obstructive jaundice liver injury in the data set. If the correlation factor is less than 0.8, the to-be-screened index is deleted; Among them, the indicator correlation factor λ α1,α2 The calculation formula is: ; Where, α 1 , α 2 They are respectively 1 , α 2 The value of Cov(α 1 ,α 2 ) is the covariance of the two indicator data, Var(α 1 )、Var(α 2 ) are their variances respectively; Among them, since in this embodiment, the prediction indicators are classified according to the percentage of occurrence of the prediction indicators in the literature, through analysis of the three types of indicators, it is found that, on the one hand, some indicators in the three types of indicators are indicators that have been included in the research in recent years, and their contribution to the prediction results is relatively high. Therefore, there may be a situation where the percentage of occurrence is relatively small. There are also some indicators that have been found through a large number of medical records to have a low contribution to the prediction results. Therefore, in the process of research, researchers gradually gave up these indicators, resulting in a situation where the percentage of occurrence is relatively small. Therefore, in this embodiment, it is necessary to screen out indicators with a high contribution to the prediction results for prediction of obstructive jaundice liver injury; specifically, as shown in the attached figure Figure 3 As shown, the three types of indicators are screened using the comprehensive screening method of predicted contribution increase value and appearance time as follows: Sa: records the year when the indicator to be screened first appears in the literature; For example, the year of publication of the use of alanine aminotransferase (ALT) content for the prediction of liver damage in obstructive jaundice is 2001, and therefore, the year in which the alanine aminotransferase (ALT) content is first recorded is 2001; Sb: Get the appearance time y of the indicator to be filtered i , where i is the number of the indicator to be screened; Among them, the occurrence time y iIt is the difference between 2004 and the year when the indicator to be screened first appears; Sc: Get the predicted contribution increase value a of the index to be screened to the predicted result i ; In the calculation process of the predicted contribution added value, all indicators in the first category are input as prediction indicators into the obstructive liver injury prediction model to predict obstructive liver injury, and the prediction accuracy of the first category of indicators is obtained; then the first category of indicators and the indicators to be screened are input as prediction indicators into the obstructive liver injury prediction model to predict obstructive liver injury, and the prediction accuracy of the first category of indicators and the indicators to be screened is obtained; the difference a between the prediction accuracy of the first category of indicators and the indicators to be screened and the prediction accuracy of the first category of indicators is calculated. i As the predicted contribution increase value of the index i to be screened; Sd: The appearance time y of the index to be screened i And the predicted contribution value of the index to be screened to the prediction result a i Screening the indicators to be screened; Specifically, the Sd is: the appearance time y of the index to be screened i The predicted contribution increase value a of the to-be-screened indicator to the predicted result is less than the first threshold value i If the value is greater than the second threshold, the indicator to be screened is retained; In this embodiment, the first threshold is 6, and the second threshold is 10%.

[0020] Among them, the final indicators for predicting obstructive jaundice liver injury after the S3 screening are procalcitonin (PCT), total bilirubin content (TBIL), γ-glutamyl transpeptidase to platelet ratio (GPR), activated partial thromboplastin time (APTT), and lactic acid content (LAO).

[0021] This embodiment first classifies the obtained prediction indicators according to the specific circumstances of the literature statistical prediction indicators into one category of indicators, two categories of indicators and three categories of indicators; the one category of indicators is not screened, the two category indicators are screened by the correlation analysis method, and the three category indicators are screened by the comprehensive screening method of the predicted contribution added value and the appearance time; the screening scheme is designed according to the specific circumstances of the prediction indicators, thereby improving the accuracy of the indicator screening; at the same time, it also avoids the situation where the accuracy is damaged due to improper selection of screening parameters caused by the use of complex model screening methods.

[0022] S4: Establish a prediction model for liver injury in obstructive jaundice; The obstructive jaundice liver injury prediction model is an artificial neural network model, a nonlinear, adaptive information processing system composed of a large number of interconnected processing units. It was proposed based on the research results of modern neuroscience and attempts to process information by simulating the way brain neural networks process and memorize information. After decades of development, artificial neural network models have achieved widespread success in many research fields, including medical prediction, pattern recognition, decision support, and artificial intelligence. Specifically, as attached Figure 4 As shown in the figure, the structure of the artificial neural network model (ANN) mainly includes three parts: neurons, layers, and networks. Neurons are the basic units of the artificial neural network model (ANN). A large number of neurons are interconnected through weights to form a complex network system. Each neuron receives multiple input signals, generates outputs through weighted summation, and then applies an activation function. The neuron structure includes dendrites (receive input signals), axons (transmit output signals), and synapses (connection points to other neurons). The artificial neural network model (ANN) is composed of multiple layers of interconnected artificial neurons, mainly divided into the following layers: the input layer receives external input data, does not perform calculations, and only passes information to the next layer; the hidden layer, located between the input and output layers, is responsible for analyzing and transforming data. It usually has multiple hidden layers, each containing multiple neurons; the output layer generates the network's final output, with each output unit corresponding to a prediction result.

[0023] S5: Obtaining a training set for training the artificial neural network model; The training set includes a normal training set and an obstructive jaundice liver injury training set. The process of obtaining the obstructive jaundice liver injury training set is as follows: searching medical records in the electronic medical record system using "obstructive jaundice" and "liver injury" as keywords, and then screening the hit medical records according to the screening rules to obtain the obstructive jaundice liver injury training set. The screening rules are: 1) Patients whose clinical records lack the final predictive indicator; 2) Patients under 18 years of age; 3) Patients who were discharged or died within 24 hours of ICU admission; 4) Patients with chronic liver disease or liver dysfunction (such as chronic hepatitis, cirrhosis, malignant liver tumors, alcoholic liver damage, and a history of liver surgery); 5) Patients with drug-induced liver damage; 6) Patients with blood system-related diseases; 7) Patients with coagulation disorders; 8) Patients with acute cardiovascular and cerebrovascular diseases; 9) Cancer patients who received radiotherapy or chemotherapy in the past three months. Patients who met any of the criteria were excluded.

[0024] S6: training the artificial neural network model using the training set of the artificial neural network model to obtain a trained artificial neural network model; Model training is the core of the neural network model. During the training process, the model needs to be iterated using the training data set, and the model parameters are continuously optimized through forward propagation and backpropagation algorithms. The forward propagation algorithm refers to inputting data into the model and calculating the model's prediction results. The backpropagation algorithm calculates the loss function based on the difference between the prediction results and the true labels, and updates the model parameters through the gradient descent algorithm. The training process adopts batch training or global training to improve training efficiency. Among them, the model's hyperparameters play a vital role in the design and training process of artificial neural network models. They are parameters set before starting training and are related to the network's structure, training process, and optimization algorithm. Accurate hyperparameters are crucial to achieving optimal model performance. Among them, the hyperparameters of the training process include batch size, number of iterations (epochs), learning rate, etc. Among them, batch size is the number of samples used to update the parameters in each iteration, the learning rate is used to control the step size of the parameter update in each iteration, and the number of iterations is used to determine the number of times the training data is completely traversed.

[0025] As a preferred embodiment, in this step, the setting of the number of iterations is improved, wherein the formula for determining the number of iterations epochs is: ; Where a and b are coefficients, L is the learning rate, and n is the number of input prediction indicators; In this embodiment, the value of a is 1, and the value of b is 10000.

[0026] As a preferred embodiment, this embodiment establishes a connection between the value of the number of iterations and the learning rate and the number of prediction indicators. When the learning rate is small, a larger number of iterations is designed. When there are more prediction indicators, a larger number of iterations is also designed. That is, the number of iterations is adjusted by the learning rate and the number of prediction indicators. In this way, the model training efficiency and training accuracy can be taken into account in the model training link.

[0027] S7: inputting the specific numerical value of the final indicator for predicting obstructive jaundice liver injury of the subject to be predicted into the trained artificial neural network model to obtain a prediction result for obstructive jaundice liver injury; The prediction result of obstructive jaundice liver injury is the probability of occurrence of obstructive jaundice liver injury.

[0028] Example 2: This example includes a system for predicting liver damage caused by obstructive jaundice. The system adopts the method for predicting liver damage caused by obstructive jaundice described in Example 1. The system includes: A prediction index acquisition module, used to acquire prediction indexes for predicting obstructive icteric liver injury; A prediction indicator classification module, used for classifying the prediction indicators; a final indicator determination module, configured to screen the prediction indicators according to the classification results of the prediction indicator classification module to obtain a final indicator for the prediction of obstructive icteric liver injury; Prediction model building module, used to establish a prediction model for obstructive jaundice liver injury; A training set acquisition module, used to acquire a training set for training the artificial neural network model; A training module, configured to train the artificial neural network model using the training set of the artificial neural network model to obtain a trained artificial neural network model; The prediction module is used to input the specific numerical value of the final indicator of the obstructive jaundice liver damage prediction of the object to be predicted into the trained artificial neural network model to obtain the obstructive jaundice liver damage prediction result.

[0029] Embodiment 3. This embodiment includes a computer-readable storage medium having a data processing program stored thereon. The data processing program is executed by a processor to implement a method for predicting liver damage caused by obstructive jaundice according to embodiment 1.

[0030] Those skilled in the art will appreciate that the embodiments herein may be provided as methods, apparatuses (devices), or computer program products. Therefore, the present invention may take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. These include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disks (DVDs) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, as is well known to those skilled in the art, communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery medium.

[0031] This document is described with reference to flowcharts and / or block diagrams of methods, apparatuses (devices), and computer program products according to the embodiments of this document. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0032] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0033] The embodiments and / or implementation methods described above are only used to illustrate the preferred embodiments and / or implementation methods for realizing the technology of the present invention, and are not intended to impose any formal restrictions on the implementation methods of the technology of the present invention. Any person skilled in the art may make some changes or modifications to other equivalent embodiments without departing from the scope of the technical means disclosed in the content of the present invention, but they should still be regarded as technologies or embodiments that are essentially the same as the present invention. Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A method for predicting liver damage caused by obstructive jaundice, characterized in that: The method comprises the following steps: S1: Obtaining predictive indicators for predicting obstructive icteric liver injury; S2: Classify the predicted indicators; The S2 is specifically: S2.1: Count the percentage of occurrences of the prediction indicators in the prediction model; S2.2: The prediction indicators are classified into first-class indicators, second-class indicators and third-class indicators according to the percentage of occurrence; S3: screening the prediction indicators according to the classification results of S2 to obtain the final indicators for predicting obstructive icteric liver injury; The S3 specifically includes: not performing a screening operation on the first type of indicators; performing a screening operation on the second type of indicators using a correlation analysis method; and performing a screening operation on the third type of indicators using a comprehensive screening method of predicted contribution increase value and appearance time; S4: Establish a prediction model for liver injury in obstructive jaundice; Wherein, the obstructive jaundice liver injury prediction model is an artificial neural network model; S5: Obtaining a training set for training the artificial neural network model; S6: training the artificial neural network model using the training set of the artificial neural network model to obtain a trained artificial neural network model; S7: Inputting the specific numerical value of the final indicator for predicting obstructive jaundice liver damage of the object to be predicted into the trained artificial neural network model to obtain a prediction result for obstructive jaundice liver damage.

2. The method for predicting liver damage caused by obstructive jaundice according to claim 1, characterized in that: In said S3, the screening operation for the two types of indicators using the correlation analysis method is specifically as follows: obtaining a data set for predicting obstructive jaundice liver damage, wherein the data set at least includes the prediction indicators in said S1 and the probability of occurrence of obstructive jaundice liver damage determined by experts based on the prediction indicators, and obtaining the correlation factor between the indicator to be screened and the probability of occurrence of obstructive jaundice liver damage in the data set; if the correlation factor is less than 0.8, deleting the indicator to be screened.

3. The method for predicting liver damage caused by obstructive jaundice according to claim 2, characterized in that: In S3, the three types of indicators are screened using a comprehensive screening method of predicted contribution increase and appearance time, specifically: Sa: records the year when the indicator to be screened first appears in the literature; Sb: Get the appearance time y of the indicator to be filtered i , where i is the number of the indicator to be screened; Sc: Get the predicted contribution increase value a of the index to be screened to the predicted result i ; Sd: The appearance time y of the index to be screened i And the predicted contribution value of the index to be screened to the prediction result a i The indicators to be screened are screened.

4. The method for predicting liver damage caused by obstructive jaundice according to claim 3, characterized in that: In the calculation process of the predicted contribution added value: all indicators in the first category are input as prediction indicators into the obstructive liver injury prediction model to predict obstructive liver injury, and the prediction accuracy of the first category of indicators is obtained; then the first category of indicators and the indicators to be screened are input as prediction indicators into the obstructive icteric liver injury prediction model to predict obstructive liver injury, and the prediction accuracy of the first category of indicators and the indicators to be screened is obtained; the difference a between the prediction accuracy of the first category of indicators and the indicators to be screened and the prediction accuracy of the first category of indicators is calculated. i As the predicted contribution increase value of the indicator i to be screened.

5. The method for predicting liver damage caused by obstructive jaundice according to claim 3, characterized in that: The Sd is specifically: the appearance time y of the index to be screened i The predicted contribution of the index to be screened to the prediction result is less than the first threshold value a i If the value is greater than the second threshold, the indicator to be screened is retained.

6. The method for predicting liver damage caused by obstructive jaundice according to claim 1, characterized in that: The final indicators for predicting obstructive icteric liver injury after the S3 screening are procalcitonin, total bilirubin content, γ-glutamyl transpeptidase to platelet ratio, activated partial thromboplastin time, and lactic acid content.

7. The method for predicting liver damage caused by obstructive jaundice according to claim 1, characterized in that: In S2.2, the first category of indicators is an indicator whose percentage of occurrence is in the interval [0.90, 1.00), the second category of indicators is an indicator whose percentage of occurrence is in the interval [0.75, 0.90), and the third category of indicators is an indicator whose percentage of occurrence is less than 0.

75.

8. The method for predicting liver damage caused by obstructive jaundice according to claim 3, characterized in that: In the Sb, the occurrence time y i It is the difference between 2004 and the year when the indicator to be screened first appears.

9. The method for predicting liver damage caused by obstructive jaundice according to claim 3, characterized in that: The Sd is specifically: the appearance time y of the index to be screened i The predicted contribution increase value a of the to-be-screened indicator to the predicted result is less than the first threshold value i If the value is greater than the second threshold, the indicator to be screened is retained.

10. A system for predicting liver damage caused by obstructive jaundice, characterized in that: The system adopts the method for predicting obstructive icteric liver injury according to any one of claims 1 to 9, and the system comprises: A prediction index acquisition module, used to acquire prediction indexes for predicting obstructive icteric liver injury; A prediction indicator classification module, used for classifying the prediction indicators; a final indicator determination module, configured to screen the prediction indicators according to the classification results of the prediction indicator classification module to obtain a final indicator for the prediction of obstructive icteric liver injury; A prediction model building module, used to build a prediction model for obstructive jaundice liver damage; wherein the prediction model for obstructive jaundice liver damage is an artificial neural network model; A training set acquisition module, used to acquire a training set for training the artificial neural network model; A training module, configured to train the artificial neural network model using the training set of the artificial neural network model to obtain a trained artificial neural network model; The prediction module is used to input the specific numerical value of the final indicator of the obstructive jaundice liver damage prediction of the object to be predicted into the trained artificial neural network model to obtain the obstructive jaundice liver damage prediction result.

Images