CN115206530A - Method and system for improving prediction precision of postoperative complications of esophageal cancer - Google Patents
Method and system for improving prediction precision of postoperative complications of esophageal cancer Download PDFInfo
- Publication number
- CN115206530A CN115206530A CN202210839836.5A CN202210839836A CN115206530A CN 115206530 A CN115206530 A CN 115206530A CN 202210839836 A CN202210839836 A CN 202210839836A CN 115206530 A CN115206530 A CN 115206530A
- Authority
- CN
- China
- Prior art keywords
- patient
- esophageal cancer
- complication
- information
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
- G16H20/60—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to nutrition control, e.g. diets
Landscapes
- Health & Medical Sciences (AREA)
- Public Health (AREA)
- Engineering & Computer Science (AREA)
- Medical Informatics (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Epidemiology (AREA)
- Data Mining & Analysis (AREA)
- Pathology (AREA)
- Databases & Information Systems (AREA)
- Biomedical Technology (AREA)
- Nutrition Science (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a method and a system for improving the prediction precision of postoperative complications of esophageal cancer, and relates to the technical field of artificial intelligence, wherein the method comprises the following steps: acquiring a patient physiological characteristic information set, a patient operation treatment information set, a patient postoperative care information set and a patient postoperative complication information set based on an esophageal cancer patient database, analyzing the patient physiological influence factor set, the operation influence factor set and the care influence factor set to obtain a physiological influence factor set, an operation influence factor set and a care influence factor set, performing factor analysis on the factor sets and the patient postoperative complication information set based on a factor regression analysis method to determine an independent risk factor set, constructing a user complication prediction model based on the independent risk factor set and the patient nutrition evaluation data information set, and performing complication prediction management on an esophageal cancer patient based on the user complication prediction model. The technical effects of improving the prediction accuracy and the prediction efficiency of the esophageal cancer complications through model fitting multi-factor evaluation are achieved.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a method and a system for improving the prediction precision of postoperative complications of esophageal cancer.
Background
Esophageal cancer is a malignant tumor that occurs in esophageal epithelial tissues and is classified into early, middle and late stages; the common treatment methods comprise operation treatment, chemotherapy treatment and drug treatment, and because the esophageal cancer operation has long time and large wound and has large influence on the respiratory, circulatory and digestive functions of patients, the postoperative complications are easy to occur. Therefore, the method can accurately predict the postoperative complications of the esophageal cancer, and has important significance for reducing the incidence rate of the complications and the treatment effect.
However, the prior art has the technical problem that the complication of the esophagus cancer is influenced by various factors, so that the prediction accuracy of the complication is low.
Disclosure of Invention
The method and the system for improving the prediction accuracy of the esophageal cancer postoperative complications solve the technical problem that the esophageal cancer complications are influenced by various factors to cause low prediction accuracy of the complications in the prior art, and achieve the technical effects of carrying out model fitting multi-factor evaluation and improving prediction accuracy and prediction efficiency of the esophageal cancer complications by acquiring physiological characteristic information, operation treatment information, postoperative care information and postoperative complication information of patients.
In view of the above problems, the present invention provides a method and system for improving accuracy of prediction of postoperative complications of esophageal cancer.
In a first aspect, the present application provides a method of improving accuracy of a prediction of postoperative complications of esophageal cancer, the method comprising: constructing an esophageal cancer patient database through big data; acquiring a patient physiological characteristic information set, a patient operation treatment information set, a patient postoperative care information set and a patient postoperative complication information set based on the esophageal cancer patient database; acquiring a physiological influence factor set, an operation influence factor set and a nursing influence factor set according to the patient physiological characteristic information set, the patient operation treatment information set and the patient postoperative nursing information set; performing factor analysis on the physiological influence factor set, the operation influence factor set, the nursing influence factor set and the patient postoperative complication information set based on a factor regression analysis method to determine an independent risk factor set; performing nutrition evaluation on each esophageal cancer patient in the esophageal cancer patient database to obtain a patient nutrition evaluation data information set; constructing a user complication prediction model based on the independent risk factor set and the patient nutrition evaluation data information set; and carrying out complication prediction management on the esophageal cancer patient based on the user complication prediction model.
In another aspect, the present application further provides a system for improving accuracy of prediction of postoperative complications of esophageal cancer, the system including: the database construction module is used for constructing an esophageal cancer patient database through big data; the patient information acquisition module is used for acquiring a patient physiological characteristic information set, a patient operation treatment information set, a patient postoperative care information set and a patient postoperative complication information set based on the esophageal cancer patient database; the influence factor acquisition module is used for acquiring a physiological influence factor set, a surgical influence factor set and a nursing influence factor set according to the patient physiological characteristic information set, the patient surgical treatment information set and the patient postoperative care information set; the factor analysis module is used for performing factor analysis on the physiological influence factor set, the operation influence factor set, the nursing influence factor set and the patient postoperative complication information set based on a factor regression analysis method to determine an independent risk factor set; the nutrition evaluation module is used for carrying out nutrition evaluation on each esophageal cancer patient in the esophageal cancer patient database to obtain a patient nutrition evaluation data information set; the model building module is used for building a user complication prediction model based on the independent risk factor set and the patient nutrition evaluation data information set; and the prediction management module is used for performing complication prediction management on the esophageal cancer patient based on the user complication prediction model.
One or more technical solutions provided in the present application have at least the following technical effects or advantages:
the method adopts the technical scheme that a patient physiological characteristic information set, a patient operation treatment information set, a patient postoperative care information set and a patient postoperative complication information set are obtained based on an esophageal cancer patient database, then a physiological influence factor set, an operation influence factor set and a care influence factor set are obtained by analyzing the physiological characteristic information set, the patient operation treatment information set and the patient postoperative complication information set, factor analysis is carried out on the physiological influence factor set, the operation influence factor set, the care influence factor set and the patient postoperative complication information set based on a factor regression analysis method, an independent risk factor set is determined, nutrition evaluation is carried out on each esophageal cancer patient simultaneously, a patient nutrition evaluation data information set is obtained, a user complication prediction model is constructed based on the independent risk factor set and the patient nutrition evaluation data information set, and complication prediction management is carried out on the esophageal cancer patient based on the user complication prediction model. And further, the technical effects of improving the accuracy and the efficiency of predicting the esophageal cancer complications by acquiring physiological characteristic information, operation treatment information, postoperative care information and postoperative complication information of a patient and performing model fitting multi-factor evaluation are achieved.
Drawings
FIG. 1 is a schematic flow chart illustrating a method for improving accuracy of esophageal cancer postoperative complication prediction according to the present application;
FIG. 2 is a schematic flow chart of the method for improving the accuracy of esophageal cancer postoperative complication prediction according to the present application for constructing an esophageal cancer patient database;
FIG. 3 is a schematic flow chart illustrating the construction of a user complication prediction model in the method for improving the accuracy of esophageal cancer postoperative complication prediction according to the present application;
FIG. 4 is a schematic diagram of a system for improving accuracy of prediction of postoperative esophageal cancer complications according to the present application;
description of the reference numerals: the system comprises a database construction module 11, a patient information acquisition module 12, an influence factor acquisition module 13, a factor analysis module 14, a nutrition evaluation module 15, a model construction module 16 and a prediction management module 17.
Detailed Description
The method and the system for improving the prediction accuracy of the esophageal cancer postoperative complications solve the technical problem that in the prior art, the esophageal cancer complications are influenced by various factors, so that the prediction accuracy of the complications is low, and achieve the technical effects of carrying out model fitting multi-factor evaluation by acquiring physiological characteristic information, operation treatment information, postoperative care information and postoperative complications information of patients, and improving the prediction accuracy and the prediction efficiency of the esophageal cancer complications.
Example one
As shown in fig. 1, the present application provides a method for improving accuracy of prediction of postoperative complications of esophageal cancer, the method comprising:
step S100: constructing an esophageal cancer patient database through big data;
as shown in fig. 2, further to construct the esophageal cancer patient database by using big data, step S100 of the present application further includes:
step S110: obtaining an esophageal cancer patient information data set through big data;
step S120: carrying out data normalization processing on the esophageal cancer patient information data set to obtain a scalar esophageal cancer patient information data set;
step S130: carrying out data conversion on the scalar esophageal cancer patient information data set according to a preset storage format to obtain a standard esophageal cancer patient information data set;
step S140: constructing a patient information label library, and performing cluster division on the standard esophageal cancer patient information data set according to the patient information label library to obtain a data cluster division result;
step S150: and carrying out classification marking on the standard esophageal cancer patient information data set based on the data clustering and partitioning result to construct the esophageal cancer patient database.
In particular, esophageal cancer is a malignant tumor that occurs in esophageal epithelial tissue, and is classified into early, middle, and late stages; the common treatment methods comprise operation treatment, chemotherapy treatment and drug treatment, and because the esophageal cancer operation has long time and large wound and has large influence on the respiratory, circulatory and digestive functions of patients, the postoperative complications are easy to occur. Therefore, the method can accurately predict the postoperative complications of the esophageal cancer, and has important significance for reducing the incidence rate and the treatment effect of the complications.
In order to construct an esophageal cancer patient database, an esophageal cancer patient information data set is obtained through big data, the esophageal cancer patient information data set is esophageal cancer patient information from multiple sources of each platform, the esophageal cancer patient information data set comprises patient names, ages, weights, physiological states, treatment information, operation information, nursing information, postoperative complication occurrence information and the like, the data size is large, the types are rich, and the database is constructed more accurately. Because all data units in the data set are not uniform, firstly, the data normalization processing is carried out on the esophageal cancer patient information data set, the normalization processing is to change a dimensional expression into a dimensionless expression, the influence caused by the non-uniform unit index dimension is eliminated, the processed scalar esophageal cancer patient information data set is obtained, and the uniform normalization of the data processing is ensured.
Data converting the scalar esophageal cancer patient information data set according to a predetermined storage format, such as data set in percent or decimal format and type: the data size format is stored, a standard esophageal cancer patient information data set is obtained, the data format is standardized, and data processing is facilitated. And constructing a patient information label library which is a patient classification label and comprises a physiological characteristic severity grade, an esophageal cancer period number, operation duration, a postoperative care grade and the like, clustering and dividing the standard esophageal cancer patient information data set according to the patient information label library, dividing the same type of data into one type, and obtaining a data clustering and dividing result.
And carrying out classification marking on the standard esophageal cancer patient information data set based on the data clustering and partitioning result, namely carrying out data label classification marking, and constructing the esophageal cancer patient database, wherein the esophageal cancer patient database is a storage database after patient data are subjected to standardization processing. Through the collection and division of patient data, the data utilization rate and the data processing standardization are improved, and the data training amount and accuracy of the subsequent prediction model construction are further guaranteed.
Step S200: acquiring a patient physiological characteristic information set, a patient operation treatment information set, a patient postoperative care information set and a patient postoperative complication information set based on the esophageal cancer patient database;
specifically, a patient physiological characteristic information set is obtained based on the esophageal cancer patient database, wherein the patient physiological characteristic information set comprises the age, the weight, the esophageal cancer tumor diameter, the tumor differentiation degree, the tumor part, the tumor period number and the like of a patient; the patient operation treatment information set comprises an operation mode, operation duration, operation intervention treatment time and the like; a patient postoperative care information set comprises postoperative care grades, care modes and the like; a set of postoperative complication information for a patient, including postoperative complication type, complication rating, time of occurrence, and the like.
Step S300: acquiring a physiological influence factor set, an operation influence factor set and a nursing influence factor set according to the patient physiological characteristic information set, the patient operation treatment information set and the patient postoperative nursing information set;
step S400: performing factor analysis on the physiological influence factor set, the operation influence factor set, the nursing influence factor set and the patient postoperative complication information set based on a factor regression analysis method to determine an independent risk factor set;
specifically, according to the patient physiological characteristic information set, the patient operation treatment information set and the patient postoperative care information set, a physiological influence factor set, an operation influence factor set and a care influence factor set are respectively extracted and obtained, wherein the influence factor sets are factors which can influence the postoperative complication of the esophageal cancer. And performing factor analysis on the physiological influence factor set, the operation influence factor set, the nursing influence factor set and the patient postoperative complication information set based on a factor regression analysis method.
The factor regression analysis method is used for quantitatively describing the linear dependence relationship between the dependent variable and the independent variable by adopting a regression equation mode so as to determine an independent risk factor set. The risk factors are factors increasing the possibility of disease or death occurrence, namely, the occurrence of the esophageal cancer complications has a certain causal relationship with the factors, and the independent risk factors are factors capable of influencing the variables, namely the esophageal cancer complications, with independent effects after eliminating the interaction among the factors in the multi-factor analysis.
Step S500: performing nutrition evaluation on each esophageal cancer patient in the esophageal cancer patient database to obtain a patient nutrition evaluation data information set;
further, in the step S500, the nutrition evaluation is performed on each esophageal cancer patient in the esophageal cancer patient database to obtain a patient nutrition evaluation data information set, where the method further includes:
step S510: obtaining a nutrition risk screening mode, and performing nutrition evaluation on each esophageal cancer patient based on the nutrition risk screening mode to obtain a nutrition evaluation information set;
step S520: performing data fitting based on the nutrition evaluation information set and the postoperative complication information set of the patient to obtain a nutrition data fitting result;
step S530: and performing significance evaluation on the fitting result of the nutritional data, and acquiring the patient nutritional evaluation data information set based on the significance evaluation result.
In particular, nutritional factors of patients with esophageal cancer are important factors affecting the occurrence of complications, and malnutrition can lead to a significant increase in the occurrence of complications. Therefore, nutrition evaluation is performed on each esophageal cancer patient in the esophageal cancer patient database based on the nutrition risk screening means, which includes european nutrition risk screening 2002 (NRS 2002), micro nutrition evaluation (MNA), subjective comprehensive evaluation (SGA), and the like, so as to obtain a nutrition evaluation information set corresponding to each means. And performing data fitting based on the nutrition evaluation information set and the postoperative complication information set of the patient to obtain a nutrition data fitting result, namely the correlation influence of the nutrition evaluation result of the patient and the postoperative complication.
And (3) performing significance evaluation on the fitting result of the nutritional data, wherein F test and t test are commonly used, and when the test result has significance, obtaining a patient nutritional evaluation data information set based on the significance evaluation result, wherein the patient nutritional evaluation data information set comprises the grade of no malnutrition, mild malnutrition, moderate malnutrition and severe malnutrition corresponding to each esophageal cancer patient. The nutrition data fitting is carried out on the patient through the nutrition evaluation tool, the accuracy of the nutrition evaluation result is improved, and the accuracy value of the complication prediction is further improved.
Step S600: constructing a user complication prediction model based on the independent risk factor set and the patient nutrition evaluation data information set;
as shown in fig. 3, further, the constructing a user complication prediction model based on the independent risk factor set and the patient nutrition evaluation data information set further includes:
step S610: identifying the independent risk factor set and the patient nutrition evaluation data information set as model sample information;
step S620: dividing the identified model sample information according to a preset proportion to obtain a training set, a verification set and a test set;
step S630: carrying out supervision training on a deep learning network model based on the training set to obtain a basic user complication prediction model;
step 640: and verifying and testing the basic user complication prediction model based on the verification set and the test set until the model prediction accuracy reaches a preset accuracy, and obtaining the user complication prediction model.
Further, step S640 of the present application further includes:
step S641: evaluating the prediction verification effect of the user complication prediction model to obtain a model output accuracy coefficient;
step S642: when the model output accuracy coefficient does not reach a preset accuracy coefficient, taking the difference value between the model output accuracy coefficient and the preset accuracy coefficient as a model optimization parameter;
step S643: and iteratively updating the user complication prediction model based on a model optimization algorithm and the model optimization parameters.
Specifically, a user complication prediction model is constructed based on the independent risk factor set and the patient nutrition evaluation data information set, and the user complication prediction model is used for performing postoperative complication prediction on the esophagus cancer patient. The independent risk factor set and the patient nutrition evaluation data information set are used as model sample information to be identified, and the identified model sample information is divided according to a predetermined proportion, illustratively, a training set, a verification set and a test set can be obtained by dividing according to the proportion of 6.
And performing supervision training on the deep learning network model based on the training set to obtain a basic user complication prediction model after data training. And verifying and testing the basic user complication prediction model based on the verification set and the test set, specifically, evaluating the prediction verification effect of the user complication prediction model to obtain a model output accuracy coefficient, wherein the larger the model output accuracy coefficient is, the higher the model output accuracy is. And when the model output accuracy coefficient does not reach a preset accuracy coefficient, indicating that the model output accuracy does not reach the standard, and taking the difference value between the model output accuracy coefficient and the preset accuracy coefficient as a model optimization parameter.
And based on a model optimization algorithm and the model optimization parameters, the common model optimization algorithm comprises a genetic algorithm, a simulated annealing algorithm, a particle swarm algorithm and the like, and the user complication prediction model is updated in an iterative manner. And (3) until the model prediction accuracy reaches a preset accuracy, wherein the preset accuracy can be set by self, for example, the preset accuracy can be set to 98%, and the user complication prediction model with the prediction accuracy reaching the standard is obtained and is used for performing postoperative complication prediction on the esophageal cancer patient. The user complication prediction model is verified and constructed through multi-source collected data, the model prediction accuracy is guaranteed, and then the complication prediction accuracy and the complication prediction efficiency are improved.
Step S700: and carrying out complication prediction management on the esophageal cancer patient based on the user complication prediction model.
Further, in the step S700 of performing complication prediction management on an esophageal cancer patient based on the user complication prediction model, the method further includes:
step S710: the user complication prediction model comprises an input layer, a regression analysis layer and an output layer;
step S720: inputting the user data information of the esophageal cancer patient into the regression analysis layer as an input layer, and outputting a user complication prediction result;
step S730: and outputting the user complication prediction result as a model output result through the output layer, and performing complication prediction management on the esophageal cancer patient based on the user complication prediction result.
Further, the method for constructing the regression analysis layer comprises the following steps:
step S711: obtaining a multiple linear regression function;
step S712: performing data fitting on the independent risk factor set and the patient nutrition evaluation data information set through the multiple linear regression function to obtain a multiple index linear regression function;
step S713: and constructing the regression analysis layer based on the multiple index linear regression function.
Specifically, the user complication prediction model includes an input layer, a regression analysis layer, and an output layer. And inputting the user data information of the esophageal cancer patient into the regression analysis layer as an input layer, wherein the regression analysis layer is used for carrying out prediction analysis on the patient data. The method for constructing the regression analysis layer specifically includes performing data fitting on the independent risk factor set and the patient nutrition evaluation data information set through a multiple linear regression function, wherein the multiple linear regression function is a function for predicting or estimating dependent variables through optimal combination of multiple independent variables, and a multiple index linear regression function after data fitting is obtained, and the multiple index linear regression function is a fitting function after regression analysis is performed on the independent risk factor set and the patient nutrition evaluation data information.
And constructing the regression analysis layer based on the multi-index linear regression function, and performing predictive fitting on user data information of the esophageal cancer patient to output a user complication prediction result, wherein the user complication prediction result comprises the complication incidence rate of the esophageal cancer patient. And outputting the user complication prediction result as a model output result through the output layer, and carrying out complication prediction management on the esophageal cancer patient based on the user complication prediction result, for example, when the complication prediction occurrence rate is too high, carrying out special care such as enteral nutrition supplement and drug treatment on the patient, reducing the complication occurrence probability, and further improving the postoperative rehabilitation effect of the patient.
In summary, the method and the system for improving the accuracy of predicting the postoperative complication of esophageal cancer provided by the present application have the following technical effects:
the method adopts the technical scheme that a patient physiological characteristic information set, a patient operation treatment information set, a patient postoperative care information set and a patient postoperative complication information set are obtained based on an esophageal cancer patient database, then a physiological influence factor set, an operation influence factor set and a care influence factor set are obtained by analyzing the physiological characteristic information set, the patient operation treatment information set and the patient postoperative complication information set, factor analysis is carried out on the physiological influence factor set, the operation influence factor set, the care influence factor set and the patient postoperative complication information set based on a factor regression analysis method, an independent risk factor set is determined, nutrition evaluation is carried out on each esophageal cancer patient simultaneously to obtain a patient nutrition evaluation data information set, a user complication prediction model is constructed based on the independent risk factor set and the patient nutrition evaluation data information set, and complication prediction management is carried out on esophageal cancer patients based on the user complication prediction model. And further, the technical effects of improving the accuracy and the efficiency of predicting the esophageal cancer complications by acquiring physiological characteristic information, operation treatment information, postoperative care information and postoperative complication information of a patient and performing model fitting multi-factor evaluation are achieved.
Example two
Based on the same inventive concept as the method for improving the accuracy of predicting the postoperative complication of esophageal cancer in the previous embodiment, the present invention further provides a system for improving the accuracy of predicting the postoperative complication of esophageal cancer, as shown in fig. 4, the system includes:
the database construction module 11 is used for constructing an esophageal cancer patient database through big data;
a patient information obtaining module 12, configured to obtain, based on the esophageal cancer patient database, a patient physiological characteristic information set, a patient surgical treatment information set, a patient postoperative care information set, and a patient postoperative complication information set;
an influence factor obtaining module 13, configured to obtain a physiological influence factor set, a surgical influence factor set, and a nursing influence factor set according to the patient physiological characteristic information set, the patient surgical treatment information set, and the patient postoperative care information set;
a factor analysis module 14, configured to perform factor analysis on the physiological influence factor set, the surgical influence factor set, the nursing influence factor set, and the post-operative complication information set of the patient based on a factor regression analysis method, and determine an independent risk factor set;
the nutrition evaluation module 15 is used for performing nutrition evaluation on each esophageal cancer patient in the esophageal cancer patient database to obtain a patient nutrition evaluation data information set;
a model construction module 16, configured to construct a user complication prediction model based on the independent risk factor set and the patient nutrition evaluation data information set;
and the prediction management module 17 is used for performing complication prediction management on the esophageal cancer patient based on the user complication prediction model.
Further, the database construction module further comprises:
the data acquisition unit is used for acquiring an esophageal cancer patient information data set through big data;
the normalization processing unit is used for carrying out data normalization processing on the esophageal cancer patient information data set to obtain a scalar esophageal cancer patient information data set;
the data conversion unit is used for carrying out data conversion on the scalar esophageal cancer patient information data set according to a preset storage format to obtain a standard esophageal cancer patient information data set;
the clustering and dividing unit is used for constructing a patient information label library, and clustering and dividing the standard esophageal cancer patient information data set according to the patient information label library to obtain a data clustering and dividing result;
and the database construction unit is used for carrying out classification marking on the standard esophageal cancer patient information data set based on the data clustering and dividing result and constructing the esophageal cancer patient database.
Further, the model building module further comprises:
the sample identification unit is used for identifying the independent risk factor set and the patient nutrition evaluation data information set as model sample information;
the sample dividing unit is used for dividing the identified model sample information according to a preset proportion to obtain a training set, a verification set and a test set;
the supervised training unit is used for carrying out supervised training on the deep learning network model based on the training set to obtain a basic user complication prediction model;
and the model verification unit is used for verifying and testing the basic user complication prediction model based on the verification set and the test set until the model prediction accuracy reaches a preset accuracy, so as to obtain the user complication prediction model.
Further, the system further comprises:
the model evaluation unit is used for evaluating the prediction verification effect of the user complication prediction model to obtain a model output accuracy coefficient;
an optimization parameter obtaining unit, configured to, when the model output accuracy coefficient does not reach a preset accuracy coefficient, take a difference between the model output accuracy coefficient and the preset accuracy coefficient as a model optimization parameter;
and the model optimization unit is used for carrying out iterative update on the user complication prediction model based on a model optimization algorithm and the model optimization parameters.
Further, the prediction management module further comprises:
the model forming unit is used for the user complication prediction model to comprise an input layer, a regression analysis layer and an output layer;
the model input unit is used for inputting the user data information of the esophageal cancer patient into the regression analysis layer as an input layer and outputting a user complication prediction result;
and the model output unit is used for outputting the user complication prediction result as a model output result through the output layer, and carrying out complication prediction management on the esophageal cancer patient based on the user complication prediction result.
Further, the model construction unit further includes:
a function obtaining unit for obtaining a multiple linear regression function;
the data fitting unit is used for performing data fitting on the independent risk factor set and the patient nutrition evaluation data information set through the multiple linear regression function to obtain a multiple index linear regression function;
and the analysis layer construction unit is used for constructing the regression analysis layer based on the multiple index linear regression function.
Further, the nutrition evaluation module further comprises:
the nutrition evaluation unit is used for obtaining a nutrition risk screening mode, and performing nutrition evaluation on each esophageal cancer patient based on the nutrition risk screening mode to obtain a nutrition evaluation information set;
the nutrition data fitting unit is used for performing data fitting on the basis of the nutrition evaluation information set and the postoperative complication information set of the patient to obtain a nutrition data fitting result;
and the significance evaluation unit is used for evaluating the significance of the fitting result of the nutritional data and obtaining the patient nutritional evaluation data information set based on the significance evaluation result.
The application provides a method for improving the accuracy of prediction of postoperative complications of esophageal cancer, which comprises the following steps: constructing an esophageal cancer patient database through big data; acquiring a patient physiological characteristic information set, a patient operation treatment information set, a patient postoperative care information set and a patient postoperative complication information set based on the esophageal cancer patient database; acquiring a physiological influence factor set, a surgical influence factor set and a nursing influence factor set according to the patient physiological characteristic information set, the patient surgical treatment information set and the patient postoperative care information set; performing factor analysis on the physiological influence factor set, the operation influence factor set, the nursing influence factor set and the patient postoperative complication information set based on a factor regression analysis method to determine an independent risk factor set; performing nutrition evaluation on each esophageal cancer patient in the esophageal cancer patient database to obtain a patient nutrition evaluation data information set; constructing a user complication prediction model based on the independent risk factor set and the patient nutrition evaluation data information set; and carrying out complication prediction management on the esophageal cancer patient based on the user complication prediction model. Solves the technical problem that the complication prediction accuracy is low because the complication of the esophagus cancer is influenced by various factors in the prior art. The method achieves the technical effects of improving the accuracy and the efficiency of predicting the esophageal cancer complications by acquiring physiological characteristic information, operation treatment information, postoperative care information and postoperative complication information of a patient and performing model fitting multi-factor evaluation.
The specification and drawings are merely illustrative of the present application, and it is intended that the present invention cover modifications and variations of this invention provided they come within the scope of the invention and their equivalents.
Claims (8)
1. A method for improving accuracy of prediction of postoperative complications of esophageal cancer, the method comprising:
constructing an esophageal cancer patient database through big data;
acquiring a patient physiological characteristic information set, a patient operation treatment information set, a patient postoperative care information set and a patient postoperative complication information set based on the esophageal cancer patient database;
acquiring a physiological influence factor set, a surgical influence factor set and a nursing influence factor set according to the patient physiological characteristic information set, the patient surgical treatment information set and the patient postoperative care information set;
performing factor analysis on the physiological influence factor set, the operation influence factor set, the nursing influence factor set and the patient postoperative complication information set based on a factor regression analysis method to determine an independent risk factor set;
performing nutrition evaluation on each esophageal cancer patient in the esophageal cancer patient database to obtain a patient nutrition evaluation data information set;
constructing a user complication prediction model based on the independent risk factor set and the patient nutrition evaluation data information set;
and carrying out complication prediction management on the esophageal cancer patient based on the user complication prediction model.
2. The method of claim 1, wherein constructing the esophageal cancer patient database from the big data comprises:
obtaining an esophageal cancer patient information data set through big data;
carrying out data normalization processing on the esophageal cancer patient information data set to obtain a scalar esophageal cancer patient information data set;
carrying out data conversion on the scalar esophageal cancer patient information data set according to a preset storage format to obtain a standard esophageal cancer patient information data set;
constructing a patient information label library, and performing cluster division on the standard esophageal cancer patient information data set according to the patient information label library to obtain a data cluster division result;
and carrying out classification marking on the standard esophageal cancer patient information data set based on the data clustering and partitioning result to construct the esophageal cancer patient database.
3. The method of claim 1, wherein constructing a user complication prediction model based on the set of independent risk factors and the set of patient nutritional assessment data information comprises:
identifying the independent risk factor set and the patient nutrition evaluation data information set as model sample information;
dividing the identified model sample information according to a preset proportion to obtain a training set, a verification set and a test set;
carrying out supervision training on the deep learning network model based on the training set to obtain a basic user complication prediction model;
and verifying and testing the basic user complication prediction model based on the verification set and the test set until the model prediction accuracy reaches a preset accuracy, and obtaining the user complication prediction model.
4. The method of claim 3, wherein the method comprises:
evaluating the prediction verification effect of the user complication prediction model to obtain a model output accuracy coefficient;
when the model output accuracy coefficient does not reach a preset accuracy coefficient, taking the difference value between the model output accuracy coefficient and the preset accuracy coefficient as a model optimization parameter;
and iteratively updating the user complication prediction model based on a model optimization algorithm and the model optimization parameters.
5. The method of claim 1, wherein said performing a predictive management of complications for esophageal cancer patients based on said user's model of complications prediction comprises:
the user complication prediction model comprises an input layer, a regression analysis layer and an output layer;
inputting the user data information of the esophageal cancer patient into the regression analysis layer as an input layer, and outputting a user complication prediction result;
and outputting the user complication prediction result as a model output result through the output layer, and performing complication prediction management on the esophageal cancer patient based on the user complication prediction result.
6. The method of claim 5, wherein the method of constructing the regression analysis layer comprises:
obtaining a multiple linear regression function;
performing data fitting on the independent risk factor set and the patient nutrition evaluation data information set through the multiple linear regression function to obtain a multiple index linear regression function;
and constructing the regression analysis layer based on the multiple index linear regression function.
7. The method of claim 1, wherein said performing a nutritional assessment of each esophageal cancer patient in said esophageal cancer patient database to obtain a set of patient nutritional assessment data information comprises:
obtaining a nutrition risk screening mode, and performing nutrition evaluation on each esophageal cancer patient based on the nutrition risk screening mode to obtain a nutrition evaluation information set;
performing data fitting based on the nutrition evaluation information set and the postoperative complication information set of the patient to obtain a nutrition data fitting result;
and performing significance evaluation on the fitting result of the nutritional data, and acquiring the patient nutritional evaluation data information set based on the significance evaluation result.
8. A system for improving accuracy of prediction of postoperative complications of esophageal cancer, the system comprising:
the database construction module is used for constructing an esophageal cancer patient database through big data;
the patient information acquisition module is used for acquiring a patient physiological characteristic information set, a patient operation treatment information set, a patient postoperative care information set and a patient postoperative complication information set based on the esophageal cancer patient database;
the influence factor acquisition module is used for acquiring a physiological influence factor set, a surgical influence factor set and a nursing influence factor set according to the patient physiological characteristic information set, the patient surgical treatment information set and the patient postoperative care information set;
a factor analysis module for performing factor analysis on the physiological influence factor set, the operation influence factor set, the nursing influence factor set and the patient postoperative complication information set based on a factor regression analysis method to determine an independent risk factor set;
the nutrition evaluation module is used for carrying out nutrition evaluation on each esophageal cancer patient in the esophageal cancer patient database to obtain a patient nutrition evaluation data information set;
the model building module is used for building a user complication prediction model based on the independent risk factor set and the patient nutrition evaluation data information set;
and the prediction management module is used for performing complication prediction management on the esophageal cancer patient based on the user complication prediction model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210839836.5A CN115206530A (en) | 2022-07-18 | 2022-07-18 | Method and system for improving prediction precision of postoperative complications of esophageal cancer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210839836.5A CN115206530A (en) | 2022-07-18 | 2022-07-18 | Method and system for improving prediction precision of postoperative complications of esophageal cancer |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115206530A true CN115206530A (en) | 2022-10-18 |
Family
ID=83582720
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210839836.5A Pending CN115206530A (en) | 2022-07-18 | 2022-07-18 | Method and system for improving prediction precision of postoperative complications of esophageal cancer |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115206530A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116825356A (en) * | 2023-07-12 | 2023-09-29 | 中国医学科学院基础医学研究所 | Multi-association surgery complication risk assessment method, system and computing equipment |
-
2022
- 2022-07-18 CN CN202210839836.5A patent/CN115206530A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116825356A (en) * | 2023-07-12 | 2023-09-29 | 中国医学科学院基础医学研究所 | Multi-association surgery complication risk assessment method, system and computing equipment |
| CN116825356B (en) * | 2023-07-12 | 2024-02-06 | 中国医学科学院基础医学研究所 | Multi-association surgery complication risk assessment method, system and computing equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109411082B (en) | Medical quality evaluation and treatment recommendation method | |
| CN109087706B (en) | Human health assessment method and system based on sleep big data | |
| US11664126B2 (en) | Clinical predictor based on multiple machine learning models | |
| US11529105B2 (en) | Digital twin updating | |
| CN110739076A (en) | medical artificial intelligence public training platform | |
| CN108492877B (en) | Cardiovascular disease auxiliary prediction method based on DS evidence theory | |
| CN113782183B (en) | Device and method for predicting risk of pressure injury based on multi-algorithm fusion | |
| Yifan et al. | Forecast model of breast cancer diagnosis based on RF-AdaBoost | |
| CN115206530A (en) | Method and system for improving prediction precision of postoperative complications of esophageal cancer | |
| CN115714022B (en) | Neonatal jaundice health management system based on artificial intelligence | |
| CN116189866A (en) | Remote medical care analysis system based on data analysis | |
| CN115205601A (en) | Medical examination result auditing system based on artificial intelligence and knowledge graph | |
| CN114943629A (en) | Health management and health care service system and health management method thereof | |
| Adi et al. | Stroke risk prediction model using machine learning | |
| CN116665843B (en) | Dietary energy intake optimization method for tumor patients | |
| US20210158967A1 (en) | Method of prediction of potential health risk | |
| CN112965971A (en) | Method for performing correlation analysis on characteristic abundance data and sample phenotype data | |
| CN111370117A (en) | Prognosis prediction system for colorectal cancer treatment population | |
| KR20210145539A (en) | Providing method for health information based on microbiome and analysis apparatus | |
| CN115547502B (en) | Hemodialysis patient risk prediction device based on time sequence data | |
| JP2024061599A (en) | A system for identifying abnormalities in the course of medical treatment based on a hierarchical neural network | |
| CN116864139A (en) | Disease risk assessment method, device, computer equipment and readable storage medium | |
| CN113593703B (en) | Device and method for constructing pressure injury risk prediction model | |
| CN112086130B (en) | Method for predicting obesity risk prediction device based on sequencing and data analysis | |
| CN114664447A (en) | Osteoporosis health risk assessment method and equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |