CN111863248A - Effective method for constructing clinical decision model - Google Patents
Effective method for constructing clinical decision model Download PDFInfo
- Publication number
- CN111863248A CN111863248A CN202010771540.5A CN202010771540A CN111863248A CN 111863248 A CN111863248 A CN 111863248A CN 202010771540 A CN202010771540 A CN 202010771540A CN 111863248 A CN111863248 A CN 111863248A
- Authority
- CN
- China
- Prior art keywords
- model
- models
- clinical decision
- training
- individuals
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000012549 training Methods 0.000 claims abstract description 39
- 230000002068 genetic effect Effects 0.000 claims abstract description 27
- 238000012795 verification Methods 0.000 claims abstract description 26
- 230000004083 survival effect Effects 0.000 claims abstract description 22
- 238000011282 treatment Methods 0.000 claims abstract description 21
- 201000010099 disease Diseases 0.000 claims abstract description 19
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims abstract description 19
- 238000012360 testing method Methods 0.000 claims abstract description 16
- 238000004458 analytical method Methods 0.000 claims abstract description 8
- 230000008901 benefit Effects 0.000 claims abstract description 8
- 230000001575 pathological effect Effects 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 238000005070 sampling Methods 0.000 claims abstract description 6
- 238000001325 log-rank test Methods 0.000 claims description 12
- 238000012216 screening Methods 0.000 claims description 11
- 238000011156 evaluation Methods 0.000 claims description 4
- 230000007170 pathology Effects 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 2
- 238000010200 validation analysis Methods 0.000 claims 2
- 230000006870 function Effects 0.000 description 11
- 230000008569 process Effects 0.000 description 11
- 208000002454 Nasopharyngeal Carcinoma Diseases 0.000 description 4
- 206010061306 Nasopharyngeal cancer Diseases 0.000 description 4
- 201000011216 nasopharynx carcinoma Diseases 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000002560 therapeutic procedure Methods 0.000 description 3
- 238000002512 chemotherapy Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000006698 induction Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000000611 regression analysis Methods 0.000 description 2
- 230000001225 therapeutic effect Effects 0.000 description 2
- 206010028980 Neoplasm Diseases 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000008280 blood Substances 0.000 description 1
- 210000004369 blood Anatomy 0.000 description 1
- 201000011510 cancer Diseases 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000013401 experimental design Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000002493 microarray Methods 0.000 description 1
- 238000011369 optimal treatment Methods 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
- 238000011277 treatment modality Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
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
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Evolutionary Computation (AREA)
- Public Health (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Artificial Intelligence (AREA)
- Genetics & Genomics (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Physiology (AREA)
- Pathology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Investigating Or Analysing Biological Materials (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses an effective method for constructing a clinical decision model, which selects a disease and a corresponding treatment means, and selects the pathological characteristics and survival outcome of a corresponding batch of patients; preprocessing data, coding pathological data, dividing the data into training samples and verification samples, establishing a clinical decision model based on a genetic algorithm, inputting corresponding disease characteristics as the characteristics of the genetic algorithm as learning samples for training, obtaining a potential model corresponding to the clinical decision of the disease, performing stability verification on the potential model corresponding to the disease, performing sampling repeated analysis on the training samples with feedback, and selecting the stable model as a candidate model. And testing the selected model by using the test sample, and taking the model passing the test as a final model. And (5) counting pathological features of the obtained final model and displaying. The method has the advantages of no need of hypothesis, small model scale, high speed and high accuracy.
Description
Technical Field
The invention relates to the technical field of medical high-dimensional feature processing, in particular to an effective method for constructing a clinical decision model.
Background
With the continuous development of computer technology, the method for combining characteristic variables in medicine is also rapidly developed, so that more and more medical data can be processed and judged by using a computer, and the success probability of disease diagnosis and treatment is improved.
However, whether a certain disease is treated by a certain therapeutic means is hardly a uniform standard in the medical field, and similar items are rarely reported. In order to obtain an optimal treatment decision model, the optimal combination of characteristic variables must be selected from a large number of characteristic variables. The high-dimensional feature space may be composed of features obtained by feature analysis. The optimal feature set is a subset of the high-dimensional feature space. Genetic algorithms have proven to be a robust global optimization algorithm for searching sub-optimal solutions in high dimensional space and for feature selection in multi-modal biometric identification systems, high dimensional cancer microarray datasets, and the like. We can see that there are many potential features available to assess whether a patient benefits from a certain therapy, and that the dimensionality of the feature space is very high. First, high-dimensional feature analysis using univariate Cox regression analysis is deficient. There are many potential models that need to be validated. It is difficult to evaluate all potential models using exhaustive methods, as this is an NP challenge. It is difficult to fully evaluate the treatment effect of a patient by a treatment decision model constructed by Cox regression analysis. Second, although LASSO can be used to construct an assessment model when there are many feature variables, the feature selection in LASSO is blind to the information whether the patient uses this therapy. In LASSO, the objective function for model construction and optimization is the deviation or area under the ROC curve (AUC), which is related only to the values of the characteristic variables involved in the analysis, and ignores the information in the feature selection using this treatment method. The LASSO-derived models are difficult to accurately make treatment decisions for patients. Third, it is difficult to know clearly whether a patient will benefit from a treatment in the clinic, i.e., it is difficult to determine a certain therapeutic effect signature for each patient. Machine learning and artificial intelligence methods can be used to build models in a high-dimensional feature space, but fail to build decision models for assessing whether a patient benefits from some treatment.
The clinical decision method for constructing the clinical decision model based on the genetic algorithm has guiding significance for whether the patient carries out the corresponding treatment method. The physician may be helped to make a more patient-appropriate treatment quicker, and combinations of characteristic variables that are more appropriate for that treatment may also be selected.
For the patient, more funds can be saved, the survival rate of the patient can be improved, and for the doctor, more reliable judgment and decision can be made aiming at the model, so that the survival rate of the patient can be increased. In summary, we introduce an event-supervised genetic algorithm for the decision-making of a patient with a certain disease using a corresponding treatment modality. The model obtained by the new method has the characteristics of simplicity and convenience for clinical application. The method provides a new and effective method for evaluating the treatment effect and treatment decision of diseases.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides an effective method for constructing a clinical decision model, which does not need to be assumed and has small model scale, high speed and high accuracy.
The technical scheme for realizing the purpose of the invention is as follows:
an effective method for constructing a clinical decision model specifically comprises the following steps:
1) preparing data: selecting a disease and a corresponding treatment means, and selecting characteristics and living data of pathology, images, clinic and the like of a corresponding batch of patients;
2) data preprocessing: preprocessing the data acquired in the step 1), namely performing binarization processing on pathological data, wherein one-hot coding is performed, a clinician performs feature screening and inputs the feature screening as binarization coding of a genetic algorithm, feature variable data of each patient are coded, and the coded patients are divided into training samples and verification samples;
3) establishing a clinical decision model based on a genetic algorithm, taking the training sample obtained in the step 2) as the characteristic input of the genetic algorithm, training the established clinical decision model, optimizing by adopting a fitness function, taking the inverse log-rank-test p value in a k-m survival curve as the fitness function, and storing all models with p less than 0.05, wherein the obtained model is a potential model for clinical decision of the corresponding disease;
4) performing stability verification on potential models of corresponding diseases, performing Bootstrap verification on the models, performing sampling repeated analysis on the potential models with returned training samples, and selecting the stable models from the potential models as candidate models;
5) inputting the verification sample into the candidate model obtained in the step 4) for testing, and obtaining a final model through testing;
6) and counting and outputting the obtained final model statistical characteristic variables.
In step 3), the training is performed by setting the following steps in the training process:
3-1) taking the reciprocal of a log-rank-test p value as a fitness function to score the individual advantages and disadvantages of the training samples in the input model, storing the two individuals with the highest score as the individuals of the next generation, performing cross exchange and variation on the remaining individuals to serve as the individuals of the next iteration, wherein the log-rank-test p is an evaluation value of the survival difference of two types of patients, and when the p is less than 0.05, the two types of patients have obvious difference, namely the survival of the patients is obviously improved, and the reciprocal of the p is taken as the fitness of a genetic algorithm to perform iteration;
3-2) initializing individuals with the number of characteristic variables exceeding a threshold value selected by the individuals in the training sample, re-randomly selecting the individuals, and defining the number of the characteristic variables of a new individual to be less than a defined threshold value;
3-3) using a method for limiting the threshold of the target population, limiting the number of screened population of the model not to be lower than the threshold number of population, and controlling the number of beneficiary population selected by the final model if the model with the number of screened population lower than the threshold number of population is not stored.
In the step 4), the Bootstrap verification is to repeatedly extract training samples in the potential models, the number of people reaches the training time of the training samples, a patient set obtained by repeated sampling is used as a verification set of the models, passing standards are specified, the passing standards are the top 2% of all test models sorted according to the passing times, and the top 2% of the models are used as candidate models.
Has the advantages that: the effective method for constructing the clinical decision model provided by the invention has the advantages of simpler model construction process, more reasonable data grouping strategy and more reasonable experimental design. The clinical application is more convenient. The model is small in scale, high in speed and high in accuracy, and applicable people who screen one treatment means aiming at different diseases provide technical support for treatment decisions of clinicians.
Drawings
FIG. 1 is a technical circuit diagram of an efficient method of constructing a clinical decision model;
FIG. 2 is a process diagram of a genetic algorithm;
FIG. 3 is a graph of a survival curve corresponding to the model in the example;
FIG. 4 is a graph of survival curves corresponding to the second model in the example;
FIG. 5 is a graph of the survival curve corresponding to the model in the example.
Detailed Description
The invention is further illustrated but not limited by the following figures and examples.
Example (b):
to test the feasibility of the method, data were provided for 719 primary diagnosed nasopharyngeal carcinoma patients using the subsidiary tumor hospital of Zhongshan university. The data contains clinical information for these nasopharyngeal carcinoma patients: survival information, accurate radiograph interpretation (MRI) information, blood information, pathology information, treatment procedure information, and the like, 533 characteristic variables in total), as test data of the present embodiment.
An effective method for constructing a nasopharyngeal carcinoma induced chemotherapy clinical decision model based on a genetic algorithm is shown in figure 1, and specifically comprises the following steps:
1) preparing data: selecting nasopharyngeal carcinoma and corresponding induced chemotherapy as treatment means, and selecting characteristics and living data of corresponding patients, such as pathology, image, clinic and the like;
2) data preprocessing: preprocessing the data acquired in the step 1), namely performing binarization processing on pathological data, wherein one-hot coding is performed, a clinician performs feature screening and inputs the feature screening as binarization coding of a genetic algorithm, feature variable data of each patient are coded, and the coded patients are divided into training samples (399 persons) and verification samples (320 persons);
3) establishing a clinical decision model based on a genetic algorithm, as shown in fig. 2, taking the training sample obtained in the step 2) as the characteristic input of the genetic algorithm, training the established clinical decision model, iterating for 1000000 generations, taking the reciprocal of the log-rank-test p value in a k-m survival curve as a fitness function, and storing all models with p less than 0.05 to obtain 72904 potential models of corresponding disease clinical decisions;
4) performing stability verification on potential models of corresponding diseases, namely performing Bootstrap verification on the models, performing sampling repeated analysis on the potential models with returned training samples, and finally selecting 1458 stable models as candidate models;
5) inputting the verification sample into the candidate model obtained in the step 4) for testing, and obtaining 3 final models through testing;
6) and counting and outputting the obtained final model statistical characteristic variables.
And 3) in the training process, the reciprocal of the log-rank-test p is used as a fitness function, the fitness function is used for evaluating the individual goodness and badness scoring standard in the genetic algorithm, the individuals with high scores are stored and used as the individuals of the next generation, the remaining individuals are subjected to cross exchange and variation and then used as the individuals of the next iteration, the log-rank-test p is an evaluation value of the survival difference of the two types of patients, when the p is less than 0.05, the two types of patients have obvious difference, namely the survival of the patients is obviously improved, and the reciprocal is used as the fitness of the genetic algorithm for iteration. In the training process, individuals with the number of the characteristic variables exceeding 8 are initialized, namely the individuals are randomized again and the number of the characteristic variables of a new individual is specified to be less than 3. And a method for limiting the threshold value of the target population is also used, the number of screened people of the model is limited to be not less than 100, and if the model less than 100 is not stored, the number of the income population selected by the final model is controlled.
In step 4), the Bootstrap verification is to repeatedly extract training samples in the potential models, so as to reach the number of people when the training samples are trained, take a patient set obtained by repeated sampling as a verification set of the models, specify a passing standard, take the passing standard as the top 2% of all test models sorted according to the passing times, and take the top 2% of the models as candidate models, namely as the candidate models of 5).
The 3 models finally obtained in step 5) are as follows:
the survival curves of the three models are shown in fig. 3, fig. 4 and fig. 5, and the graphs show that the suggested induction population and the non-suggested induction population are obviously different, the p values are all less than 0.05, the model analysis effect is good, and the correctness of the method is proved.
In the embodiment, a clinical decision model screening mechanism based on a genetic algorithm is established, the classical genetic algorithm is referred, a conventional fitness function in the genetic algorithm is optimized, the reciprocal of log-rank-test p in a k-m survival curve is taken as the fitness function, the training process is directly supervised, so that the resolvable performance is enhanced, the accuracy and the credibility are greatly enhanced, and all models with the p less than 0.05 are stored. And (5) performing stability verification and test sample verification on the model with the p value less than 0.05, and screening the model again.
The model screened by the genetic algorithm is obtained by training survival of a certain disease aiming at a certain treatment means, and characteristic variables of the population recommended to be treated can be combined;
the innovation of the invention is that the reciprocal of log-rank-test p is adopted as a fitness function, the fitness function is a standard for evaluating the quality and the grade of individuals in a genetic algorithm, two individuals with the highest grade are stored as the individuals of the next generation, the rest individuals are subjected to cross exchange and variation to be used as the individuals of the next iteration, the log-rank-test p is an evaluation value of the survival difference of two types of patients, when p is less than 0.05, the two types of patients have obvious difference, namely the survival of the patients is obviously improved, the reciprocal is used as the fitness of the genetic algorithm for iteration, the applicability of the medical field is greatly improved, only the optimal model is reserved without engineering, and all models with p less than 0.05 are stored, namely all models with medical significance are stored, so that the problem of trapping in a local optimal solution caused by small data volume is greatly avoided. And in the training process, initializing the individuals of which the number of the characteristic variables selected by the individuals exceeds the threshold, namely, re-randomly selecting the individuals and specifying that the number of the characteristic variables of a new individual is less than the specified threshold, calling back the number of the selected combined variables, adjusting the number of the selected characteristic variables, refining the area to be searched and searching for multiple times, so that more effective models are searched, and the condition that the search area is meaningless due to precocity is avoided. And a method for limiting the threshold value of the target population is also used in the training process, the number of people screened by the model is limited to be not less than the threshold number of people, if the model with the number of people less than the threshold number of people is not stored, the number of people of the income population selected by the final model is controlled, the log-rank-testp value is prevented from being meaningless due to too few people, and the ideal beneficial population range is controlled. By means of the Bootstrap verification method for the model, repeated extraction is carried out on a returned training sample to reach the number of people when the training sample is trained, the number of people is used as a verification set of the model, a threshold value (percentage of verification number) is specified, the model with the verification passing frequency exceeding the threshold value is used as a candidate model, namely the model is subjected to Bootstrap verification, the model meeting requirements due to the fact that data accidentally occur during training is screened out, the stability of the model is verified, and the repeatable verifiability of the model is greatly increased; the model is externally verified through the verification of a test sample, and the performance of the model in different data samples is verified again, so that the reliability of the model is greatly improved; the fact proves that the model screening method is also effective, and the characteristic combination of the patient suitable for a certain therapy is accurately obtained, so that the survival rate of the patient is improved, and unnecessary waste of medical resources is reduced.
Most of the processes are model screening processes, and in actual operation, only pathological data of a patient with a certain disease is used, and then a characteristic variable combination suitable for the patient with a corresponding treatment method can be obtained;
the above is a detailed technical principle of the present invention, in which an improved genetic algorithm and a specific screening process after obtaining a model are introduced in detail.
Claims (3)
1. An efficient method for constructing a clinical decision model, comprising the steps of:
1) preparing data: selecting a disease and a corresponding treatment means, and selecting characteristics and living data of pathology, images, clinic and the like of a corresponding batch of patients;
2) data preprocessing: preprocessing the data acquired in the step 1), namely performing binarization processing on pathological data, wherein one-hot coding is performed, a clinician performs feature screening and inputs the feature screening as binarization coding of a genetic algorithm, feature variable data of each patient are coded, and the coded patients are divided into training samples and verification samples;
3) establishing a clinical decision model based on a genetic algorithm, taking the training sample obtained in the step 2) as the characteristic input of the genetic algorithm, training the established clinical decision model, optimizing by adopting a fitness function, taking the inverse log-rank-test p value in a k-m survival curve as the fitness function, and storing all models with p less than 0.05, wherein the obtained model is a potential model for clinical decision of the corresponding disease;
4) performing stability verification on potential models of corresponding diseases, performing Bootstrap verification on the models, performing sampling repeated analysis on the potential models with returned training samples, and selecting the stable models from the potential models as candidate models;
5) inputting the verification sample into the candidate model obtained in the step 4) for testing, and obtaining a final model through testing;
6) and counting and outputting the obtained final model statistical characteristic variables.
2. An efficient method for constructing a clinical decision model according to claim 1, wherein in step 3), the training is performed by setting the following settings during the training:
2-1) taking the reciprocal of a log-rank-test p value as a fitness function to score the individual advantages and disadvantages of the training samples in the input model, storing the two individuals with the highest score as the individuals of the next generation, performing cross exchange and variation on the remaining individuals to serve as the individuals of the next iteration, wherein the log-rank-test p is an evaluation value of the survival difference of two types of patients, and when the p is less than 0.05, the two types of patients have obvious difference, namely the survival of the patients is obviously improved, and the reciprocal of the p is taken as the fitness of a genetic algorithm to perform iteration;
3-2) initializing individuals with the number of characteristic variables exceeding a threshold value selected by the individuals in the training sample, re-randomly selecting the individuals, and defining the number of the characteristic variables of a new individual to be less than a defined threshold value;
3-3) using a method for limiting the threshold of the target population, limiting the number of screened population of the model not to be lower than the threshold number of population, and controlling the number of beneficiary population selected by the final model if the model with the number of screened population lower than the threshold number of population is not stored.
3. An efficient method for constructing a clinical decision model according to claim 1, wherein in step 4), Bootstrap validation is repeated extraction of training samples in potential models, the number of people when training samples are obtained, the repeatedly extracted patient set is used as the validation set of models, passing criteria is defined, the criterion for passing is the top 2% of all test models in the order of passing, and the top 2% of models are used as candidate models.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010771540.5A CN111863248B (en) | 2020-08-04 | 2020-08-04 | Effective method for constructing clinical decision model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010771540.5A CN111863248B (en) | 2020-08-04 | 2020-08-04 | Effective method for constructing clinical decision model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111863248A true CN111863248A (en) | 2020-10-30 |
| CN111863248B CN111863248B (en) | 2022-10-11 |
Family
ID=72953184
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010771540.5A Expired - Fee Related CN111863248B (en) | 2020-08-04 | 2020-08-04 | Effective method for constructing clinical decision model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111863248B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114121204A (en) * | 2021-12-09 | 2022-03-01 | 上海森亿医疗科技有限公司 | Patient record matching method based on patient master index, storage medium and equipment |
| CN114334161A (en) * | 2021-12-30 | 2022-04-12 | 医渡云（北京）技术有限公司 | Model training method, data processing method, device, medium and electronic equipment |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104867074A (en) * | 2015-05-15 | 2015-08-26 | 东北大学 | Student comprehensive quality evaluation method based on genetic algorithm optimization BP neural network |
| CN107273717A (en) * | 2017-05-27 | 2017-10-20 | 中国人民解放军第三0七医院 | A kind of detection model of Sera of Lung Cancer gene and its construction method and application |
| US20170316153A1 (en) * | 2009-10-29 | 2017-11-02 | True Health Ip Llc | Method for determining risk of diabetes |
| CN109346177A (en) * | 2018-09-10 | 2019-02-15 | 昆明理工大学 | A kind of disease symptoms quantitative forecasting technique based on improvement GA-BP algorithm |
| CN110031624A (en) * | 2019-02-28 | 2019-07-19 | 中国科学院上海高等研究院 | Tumor markers detection system based on multiple neural networks classifier, method, terminal, medium |
| CN110459328A (en) * | 2019-07-05 | 2019-11-15 | 梁俊 | A kind of Clinical Decision Support Systems for assessing sudden cardiac arrest |
| CN111128328A (en) * | 2019-10-25 | 2020-05-08 | 中山大学肿瘤防治中心（中山大学附属肿瘤医院、中山大学肿瘤研究所） | Nasopharyngeal carcinoma structured image report and data processing system and method |
| CN111139300A (en) * | 2020-02-19 | 2020-05-12 | 伯克利南京医学研究有限责任公司 | Application of group of colon cancer prognosis related genes |
-
2020
- 2020-08-04 CN CN202010771540.5A patent/CN111863248B/en not_active Expired - Fee Related
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170316153A1 (en) * | 2009-10-29 | 2017-11-02 | True Health Ip Llc | Method for determining risk of diabetes |
| CN104867074A (en) * | 2015-05-15 | 2015-08-26 | 东北大学 | Student comprehensive quality evaluation method based on genetic algorithm optimization BP neural network |
| CN107273717A (en) * | 2017-05-27 | 2017-10-20 | 中国人民解放军第三0七医院 | A kind of detection model of Sera of Lung Cancer gene and its construction method and application |
| CN109346177A (en) * | 2018-09-10 | 2019-02-15 | 昆明理工大学 | A kind of disease symptoms quantitative forecasting technique based on improvement GA-BP algorithm |
| CN110031624A (en) * | 2019-02-28 | 2019-07-19 | 中国科学院上海高等研究院 | Tumor markers detection system based on multiple neural networks classifier, method, terminal, medium |
| CN110459328A (en) * | 2019-07-05 | 2019-11-15 | 梁俊 | A kind of Clinical Decision Support Systems for assessing sudden cardiac arrest |
| CN111128328A (en) * | 2019-10-25 | 2020-05-08 | 中山大学肿瘤防治中心（中山大学附属肿瘤医院、中山大学肿瘤研究所） | Nasopharyngeal carcinoma structured image report and data processing system and method |
| CN111139300A (en) * | 2020-02-19 | 2020-05-12 | 伯克利南京医学研究有限责任公司 | Application of group of colon cancer prognosis related genes |
Non-Patent Citations (6)
| Title |
|---|
| N. K. CHLIS等: "A generic framework for the elicitation of stable and reliable gene expression signatures", 《13TH IEEE INTERNATIONAL CONFERENCE ON BIOINFORMATICS AND BIOENGINEERING》 * |
| 吴莉青等: "基于磁共振成像及调强放疗的鼻咽癌副鼻窦侵犯的预后分析", 《临床放射学杂志》 * |
| 杨青: "基于改进贝叶斯网络的乳腺癌预后模型的研究与应用", 《中国优秀硕士学位论文全文数据库 (医药卫生科技辑)》 * |
| 武海滨等: "基于偏最小二乘线性判别分析的遗传算法在代谢组学特征筛选中的应用", 《中国卫生统计》 * |
| 王宇燕等: "改进随机森林的集成分类方法预测结直肠癌存活性", 《管理科学》 * |
| 袁锋等: "一种冠心病介入术后复发预测组合模型的构建", 《计算机工程与应用》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114121204A (en) * | 2021-12-09 | 2022-03-01 | 上海森亿医疗科技有限公司 | Patient record matching method based on patient master index, storage medium and equipment |
| CN114334161A (en) * | 2021-12-30 | 2022-04-12 | 医渡云（北京）技术有限公司 | Model training method, data processing method, device, medium and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111863248B (en) | 2022-10-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | Automatic detection of diabetic retinopathy in retinal fundus photographs based on deep learning algorithm | |
| Alzu’bi et al. | Kidney tumor detection and classification based on deep learning approaches: a new dataset in CT scans | |
| Gull et al. | Artificial intelligence in brain tumor detection through MRI scans: advancements and challenges | |
| Walsh et al. | Deep learning–based outcome prediction in progressive fibrotic lung disease using high-resolution computed tomography | |
| Nahiduzzaman et al. | Hybrid CNN-SVD based prominent feature extraction and selection for grading diabetic retinopathy using extreme learning machine algorithm | |
| CN111758108A (en) | System and method for modeling probability distributions | |
| Obulesu et al. | [Retracted] Adaptive Diagnosis of Lung Cancer by Deep Learning Classification Using Wilcoxon Gain and Generator | |
| CN113113130A (en) | Tumor individualized diagnosis and treatment scheme recommendation method | |
| CN111863248B (en) | Effective method for constructing clinical decision model | |
| Bruno et al. | Using heatmaps for deep learning based disease classification | |
| Nedra et al. | Detection and classification of the breast abnormalities in Digital Mammograms via Linear Support Vector Machine | |
| Reddy et al. | Intelligent deep learning algorithm for lung cancer detection and classification | |
| CN116259415A (en) | Patient medicine taking compliance prediction method based on machine learning | |
| Singh et al. | An analysis of detection and diagnosis of different classes of skin diseases using artificial intelligence-based learning approaches with hyper parameters | |
| Challab et al. | Ant colony optimization–rain optimization algorithm based on hybrid deep learning for diagnosis of lung involvement in coronavirus patients | |
| Vinoth et al. | Pneumonia Detection from Chest X-Ray using AlexNet Image Classification Technique | |
| Dakdareh et al. | Diagnosis of Alzheimer’s Disease and Mild Cognitive Impairment Using Convolutional Neural Networks | |
| Abdulrahman et al. | Classification of retina diseases from OCT using genetic programming | |
| Madhu et al. | Covid-19 Classification and Detection Model using Deep Learning | |
| Ayoola et al. | A Comparative Analysis of Regression Algorithms with Genetic Algorithm In The Prediction of Breast Cancer Tumors | |
| Godbin et al. | Leveraging Radiomics and Genetic Algorithms to Improve Lung Infection Diagnosis in X-Ray Images using Machine Learning | |
| Prabhavathi et al. | Analysis of Lung Tissue Densities for Lung Cancer Classification Using Deep Learning Methodologies | |
| Bolourchi et al. | A machine learning-based data-driven approach to Alzheimer’s disease diagnosis using statistical and harmony search methods | |
| Graselin et al. | A Systematic Review based on the Detection of PCOS using Machine Learning Techniques | |
| Liu et al. | A batch normalization autoencoder model for breast cancer multidimensional follow-up data |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20221011 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |