CN110031624A - Tumor markers detection system based on multiple neural networks classifier, method, terminal, medium - Google Patents
Tumor markers detection system based on multiple neural networks classifier, method, terminal, medium Download PDFInfo
- Publication number
- CN110031624A CN110031624A CN201910149298.5A CN201910149298A CN110031624A CN 110031624 A CN110031624 A CN 110031624A CN 201910149298 A CN201910149298 A CN 201910149298A CN 110031624 A CN110031624 A CN 110031624A
- Authority
- CN
- China
- Prior art keywords
- data
- sample
- sample data
- model
- abnormal
- 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
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/48—Biological material, e.g. blood, urine; Haemocytometers
- G01N33/50—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing
- G01N33/53—Immunoassay; Biospecific binding assay; Materials therefor
- G01N33/574—Immunoassay; Biospecific binding assay; Materials therefor for cancer
- G01N33/57407—Specifically defined cancers
- G01N33/57438—Specifically defined cancers of liver, pancreas or kidney
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Theoretical Computer Science (AREA)
- Immunology (AREA)
- Urology & Nephrology (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Chemical & Material Sciences (AREA)
- Biophysics (AREA)
- Hematology (AREA)
- Computing Systems (AREA)
- Oncology (AREA)
- Biotechnology (AREA)
- Cell Biology (AREA)
- Hospice & Palliative Care (AREA)
- Microbiology (AREA)
- Gastroenterology & Hepatology (AREA)
- Food Science & Technology (AREA)
- Medicinal Chemistry (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Pathology (AREA)
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The application provides tumor markers detection system, method, terminal, medium based on multiple neural networks classifier, the main data preprocessing method for using principal component analysis and feature extraction, based on random forest, support vector machines, BP neural network, and the assembled classifier model of extreme learning machine sorter model is to data training, finally obtain an accuracy, specificity, sensitivity meets the hepatic carcinoma marker classifier of clinical diagnosis, it helps clinician to reduce misdiagnosis rate in liver cancer first visit, improves the accuracy of hepatic carcinoma marker detection.
Description
Technical field
This application involves neural network classifier technical fields, more particularly to the tumour based on multiple neural networks classifier
Marker detection system, method, terminal, medium.
Background technique
Primary carcinoma of liver is the 4th common cancer in current China and the 3rd tumor lethal cause of disease, is seriously threatened
The life and health of our people.The detection means of hepatocarcinoma early diagnosis is broadly divided into two major classes at present:
One is using liver cancer serum tumor markers as the detection means of reference index.Newest " the primary carcinoma of liver in China
Diagnosis and treatment specification 2017 editions " in point out, hepatic carcinoma marker AFP AFP is that Current Diagnostic liver cancer is common and important side
Method, clinically, when the content of AFP AFP is greater than 400ug/L, there may be liver cancer for prompt.Although diagnosis of the AFP to liver cancer
Sensibility and specificity with higher, but still there is 40% early liver cancer and the advanced liver cancer patient of 15%-20% may occur in which
False positive.Another kind is with ultrasound, computer X-ray, Magnetic resonance imaging, digital subtraction angiography and hepatopathy reason puncture etc.
Detection technique is the iconography of representative, pathology means.
But former detection means depends on the practical work experience of doctor, needs by long-term practice
Outstanding level of decision-making can be reached, by the subjective experience of doctor and being affected for external interference factor；And latter detects hand
Most of technologies in section are limited in the early diagnosis ability of liver cancer, not only have certain influence, but also price to the body of patient
Height does not have universality.
There are many type of liver cancer serum tumor markers, but clinically can make a definite diagnosis liver without any single tumor markers
Cancer, the presence of each tumor markers have certain reference value in diagnosis, but also have the limitation of itself.It is clinical
On, doctor carries out joint-detection diagnosis referring generally to the parameter of Diagnostic Value of Several Serum Tumor Markers.The factor of artificial experience, which is be easy to cause, examines
It will appear error on disconnected.The accuracy of Diagnostic Value of Several Serum Tumor Markers Combining diagnosis can be effectively improved using machine learning means.
Computer-aided diagnosis algorithm based on machine learning is continued to optimize and perfect in recent years, this is multi-tumor marker Combining diagnosis
The building of model brings possibility.In terms of algorithm, foreign countries have scholar to compare 179 different classifications devices in 121 different numbers
According to the actual effect on collection.Result of study shows: effect of the different classifications device under different classifications scene is different, such as with
Machine forest RF is most strong on average, but has also only taken first on 9.9% data set；The average water of support vector machines
Followed by flat, first is taken on 10.7% data set.
Therefore, this field needs a kind of technical side of accuracy that can effectively improve Diagnostic Value of Several Serum Tumor Markers Combining diagnosis
Case.
Apply for content
In view of the foregoing deficiencies of prior art, the application is designed to provide based on multiple neural networks classifier
Tumor markers detection system, method, terminal, medium, for solving the problems of the prior art.
In order to achieve the above objects and other related objects, the first aspect of the application provides a kind of based on multiple neural network point
The tumor markers detection system of class device comprising: sample collection module, for acquiring the sample of tumor-marker analyte detection sample
Data, wherein the sample data includes test set data and training set data；Data preprocessing module, for according to the sample
Notebook data filters out multiple abnormal indexes with tumour with high relevance；Sample statistics analysis module, for being based on the sample
Notebook data data do testing result analysis and assessment to single abnormal index, and/or do detection knot to multiple united abnormal indexes
Fruit analysis and assessment；Sorter model training module, for carrying out model to multiple sorter models based on the training set data
Training；Sorter model assessment and test module, for being carried out based on the test set data to the sorter model after training
Test, and compared with test result is done superiority and inferiority with the analysis and assessment result of the sample statistics analysis module, judgement is tested accordingly
The validity of the sorter model of examination；Classifier diagnostic method application module is judged as effective classifier mould for utilizing
Type detects tumor markers data, exports diagnostic result information accordingly.
In some embodiments of the first aspect of the application, the clinical true sample number of sample collection module acquisition
Classify according to and to sample data collected.
In some embodiments of the first aspect of the application, the tumor markers include hepatic carcinoma marker；Its
In, the mode classification of the sample data includes: that sample data is divided into liver cancer group sample data and non-liver cancer group sample data,
And/or it is divided into liver cancer group sample data, hepatopathy group sample data and health group sample data.
In some embodiments of the first aspect of the application, the data preprocessing module is used to be based on principal component analysis
Algorithm carries out feature selecting, to carry out dimension-reduction treatment to the multiple abnormal indexes gone out according to the sample data initial screening, and
Obtain the abnormal index with tumour with high relevance.
It is described to include with tumour has high relevance abnormal index in some embodiments of the first aspect of the application
With hepatic carcinoma have high relevance abnormal index comprising alpha-fetoprotein, further includes: abnormal prothrombin, carcinomebryonic antigen,
Carbohydrate antigen 199, carbohydrate antigen 242, sugar antigen 211, carbohydrate antigen 125, sialic acid or ferritin.
In some embodiments of the first aspect of the application, the evaluation index of testing result analysis and assessment includes sensitive
Degree, specificity and accuracy；Wherein, sensitivity indicates the practical ratio to be judged as the positive in positive sample, specificity table
Show the practical ratio to be judged as feminine gender in negative sample, accuracy indicates that true positives and true negative number account for tested total number of persons
Ratio.
In some embodiments of the first aspect of the application, the sorter model includes: Random Forest model, supports
Vector machine model, BP neural network model, extreme learning machine model.
In some embodiments of the first aspect of the application, the sample statistics analysis module is according to ROC curve from list
Selected part or the abnormal index of whole are as united abnormal index in a abnormal index.
In order to achieve the above objects and other related objects, the second aspect of the application provides a kind of based on multiple neural network point
The tumor-marker object detecting method of class device comprising: the sample data of acquisition tumor-marker analyte detection sample, wherein the sample
Data include test set data and training set data；Filtering out multiple and tumour according to the sample data has high relevance
Abnormal index；Testing result analysis and assessment are done to single abnormal index based on the sample data data, and/or to multiple joints
Abnormal index do testing result analysis and assessment, to generate corresponding analysis and assessment result；Based on the training set data to more
A sorter model carries out model training；The sorter model after training is tested based on the test set data, and will
Compared with test result does superiority and inferiority with the analysis and assessment result, the validity of tested sorter model is judged accordingly；It utilizes
It is judged as effective sorter model to detect tumor markers data, exports diagnostic result information accordingly.
In order to achieve the above objects and other related objects, the third aspect of the application provides a kind of computer-readable storage medium
Matter is stored thereon with computer program, realizes when the computer program is executed by processor described based on multiple neural network point
The tumor-marker object detecting method of class device.
In order to achieve the above objects and other related objects, the fourth aspect of the application provides a kind of detection terminal, comprising: place
Manage device and memory；The memory is used to execute the memory storage for storing computer program, the processor
Computer program, so that the detection terminal executes the tumor-marker object detecting method based on multiple neural networks classifier.
As described above, the tumor markers detection system based on multiple neural networks classifier of the application, method, terminal,
Medium has the advantages that the application provides the hepatic carcinoma marker detection scheme based on multiple neural networks classifier,
Using the data preprocessing method of principal component analysis and feature extraction, it is based on random forest, support vector machines, BP neural network,
And the assembled classifier model of extreme learning machine sorter model finally obtains an accuracy to data training, specificity,
Sensitivity meets the hepatic carcinoma marker classifier of clinical diagnosis, and clinician is helped to reduce misdiagnosis rate in liver cancer first visit,
Improve the accuracy of hepatic carcinoma marker detection.
Detailed description of the invention
Fig. 1 is shown as showing for the tumor markers detection system in one embodiment of the application based on multiple neural networks classifier
It is intended to.
Fig. 2 is shown as the ROC curve diagram drawn in one embodiment of the application based on 6 kinds of hepatic carcinoma markers.
Fig. 3 is shown as showing for the tumor-marker object detecting method in one embodiment of the application based on multiple neural networks classifier
It is intended to.
Fig. 4 is shown as detecting the structural schematic diagram of terminal in one embodiment of the application.
Specific embodiment
Illustrate presently filed embodiment below by way of specific specific example, those skilled in the art can be by this specification
Disclosed content understands other advantages and effect of the application easily.The application can also pass through in addition different specific realities
The mode of applying is embodied or practiced, the various details in this specification can also based on different viewpoints and application, without departing from
Various modifications or alterations are carried out under spirit herein.It should be noted that in the absence of conflict, following embodiment and implementation
Feature in example can be combined with each other.
It should be noted that with reference to attached drawing, attached drawing describes several embodiments of the application in described below.It should
Understand, other embodiments also can be used, and mechanical group can be carried out without departing substantially from spirit and scope
At, structure, electrical and operational change.Following detailed description should not be considered limiting, and the application
The range of embodiment only limited by the claims for the patent announced.Term used herein is merely to description is specific
Embodiment, and it is not intended to limit the application.The term of space correlation, for example, "upper", "lower", "left", "right", " following ", " under
Side ", " lower part ", " top ", " top " etc. can be used in the text in order to elements or features shown in explanatory diagram and another
The relationship of one elements or features.
In this application unless specifically defined or limited otherwise, term " installation ", " connected ", " connection ", " fixation ",
Terms such as " fixings " shall be understood in a broad sense, for example, it may be being fixedly connected, may be a detachable connection, or be integrally connected；
It can be mechanical connection, be also possible to be electrically connected；It can be directly connected, can also indirectly connected through an intermediary, it can be with
It is the connection inside two elements.For the ordinary skill in the art, above-mentioned art can be understood as the case may be
The concrete meaning of language in this application.
Furthermore as used in herein, singular " one ", "one" and "the" are intended to also include plural number shape
Formula, unless there is opposite instruction in context.It will be further understood that term "comprising", " comprising " show that there are the spies
Sign, operation, element, component, project, type, and/or group, but it is not excluded for one or more other features, operation, element, group
Presence, appearance or the addition of part, project, type, and/or group.Term "or" and "and/or" used herein are interpreted as including
Property, or mean any one or any combination.Therefore, " A, B or C " or " A, B and/or C " mean " it is following any one:
A；B；C；A and B；A and C；B and C；A, B and C ".Only when the combination of element, functions or operations is inherently mutual under certain modes
When repulsion, it just will appear the exception of this definition.
There are many type of liver cancer serum tumor markers, but clinically can make a definite diagnosis liver without any single tumor markers
Cancer, the presence of each tumor markers have certain reference value in diagnosis, but also have the limitation of itself.It is clinical
On, doctor carries out joint-detection diagnosis referring generally to the parameter of Diagnostic Value of Several Serum Tumor Markers.The factor of artificial experience, which is be easy to cause, examines
It will appear error on disconnected.The accuracy of Diagnostic Value of Several Serum Tumor Markers Combining diagnosis can be effectively improved using machine learning means.
Computer-aided diagnosis algorithm based on machine learning is continued to optimize and perfect in recent years, this is multi-tumor marker Combining diagnosis
The building of model brings possibility.In terms of algorithm, foreign countries have scholar to compare 179 different classifications devices in 121 different numbers
According to the actual effect on collection.Result of study shows: effect of the different classifications device under different classifications scene is different, such as with
Machine forest RF is most strong on average, but has also only taken first on 9.9% data set；The average water of support vector machines
Followed by flat, first is taken on 10.7% data set.
In view of above-mentioned various technical problems, the application provides the tumor markers detection system based on multiple neural networks classifier
System, method, terminal, medium, effectively to solve those problems.The main thought of the application is intended to provide based on multiple neural network point
The hepatic carcinoma marker detection scheme of class device, using the data preprocessing method of principal component analysis and feature extraction, based on
The assembled classifier model of machine forest, support vector machines, BP neural network and extreme learning machine sorter model instructs data
Practice, finally obtain an accuracy, specificity, sensitivity meets the hepatic carcinoma marker classifier of clinical diagnosis, helps to face
Bed doctor reduces misdiagnosis rate in liver cancer first visit, improves the accuracy of hepatic carcinoma marker detection.
As shown in Figure 1, showing the tumor markers detection system based on multiple neural networks classifier in one embodiment of the application
The schematic diagram of system.The system comprises sample acquisition module 11, data preprocessing module 12, sample statistics analysis module 13, divide
Class device model training module 14, sorter model assessment and test module 15, classifier diagnostic method application module 16.
Sample acquisition module 11 is used to acquire the sample data of tumor-marker analyte detection sample.
It in one embodiment, is the training effect for ensuring system, hepatic carcinoma marker detection sample preferably satisfies data
Source is truthful data and the requirement classified to sample data.Specifically, hepatic carcinoma marker detection samples sources
In realistic case, compared to the data expanded by machine learning means, it can be ensured that the testing result of detection system it is true
Reality.In addition, hepatic carcinoma marker detection sample can be divided into liver cancer group and non-liver cancer group two major classes, or liver cancer can be divided into
Group, hepatopathy group and health group are these three types of.
It is worth noting that, need to guarantee that every group of sample size cannot be too low, the sample size between group and group is protected as far as possible
Hold the same order of magnitude.Meanwhile sample is divided into training set and test set, the ratio of training set and test set is preferably 7:3.It surveys
It requires in examination collection and training set comprising sample all categories.
Data preprocessing module 12, which is used to filter out multiple and hepatic carcinoma according to the sample data, has high relevance
Abnormal index.
In one embodiment, data preprocessing module first rejects the category unrelated with subject before filtering out abnormal index
Property, to promote the accuracy of the selection result.In this present embodiment, it is simultaneously final true to perform multi-turns screen for data preprocessing module
Fixed and hepatic carcinoma has 6 abnormal indexes of high relevance.
Specifically, first filtering out the abnormal index more than 20 kinds after rejecting unrelated attribute data.Utilize Multiple Imputation
Data are carried out with the filling of missing values, and utilizes Algorithm of Mining Association Rules, from 20 kinds or more of the abnormal index previously screened
In filter out with strongest about 10 kinds of the abnormal index of primary carcinoma of liver relevance, finally utilize Principal Component Analysis Algorithm progress
Feature selecting is to carry out dimension-reduction treatment to data.
In this present embodiment, the hepatic carcinoma marker that finishing screen is selected has: alpha-fetoprotein (AFP), other auxiliary references
Tumor markers have abnormal prothrombin (PIVKA-II), carcinomebryonic antigen (CEA)), carbohydrate antigen 199 (CA199), carbohydrate is anti-
242 (CA242) of original, sugar antigen 211 (CA211), carbohydrate antigen 125, sialic acid, ferritin etc..
It is replaced it should be noted that Multiple Imputation described in the present embodiment refers to by the vector comprising multiple interpolation values
A series of process of each missing values, that is, replacing each missing values with possible values.Excavation in the present embodiment is closed
Connection rule-based algorithm is, for example, Apriori algorithm, and core concept is to detect two by the downward closing of candidate generation and plot
A stage carrys out Mining Frequent Itemsets Based.
Sample statistics analysis module 13 is used to do testing result to single abnormal index based on the sample data data and comment
Estimate analysis, and/or testing result analysis and assessment are done to multiple united abnormal indexes.
In one embodiment, the tumor markers sample data chosen is carried out respectively using statistics software spss12.0
Descriptive analysis.Descriptive analysis, which refers to, analyzes collected data, show that the various quantity of reflection objective phenomenon are special
A kind of analysis method of sign, it includes the central tendency analysis of data, data discrete degree analyzing, the analysis of frequency distribution of data
Etc..
Specifically, being analyzed respectively single tumor-marker analyte detection and joint-detection.Using non-parametric test come
Judge whether the difference between different groups has statistical significance, is test stone with α=0.05.With diagnostic sensitivity, specifically
Degree, accuracy and ROC curve are as evaluation index.Calculation formula formula 1 as follows)~3) shown in:
Sensitivity TPR=TP/ (TP+FN) × 100%；Formula 1)
Specificity TNR=TN/ (TN+FP) × 100%；Formula 2)
Accuracy=(TP+TN)/(TP+TN+FP+FN) × 100%；Formula 3)
Wherein, sensitivity refers to practical to be judged as positive ratio in positive sample that specificity refers to practical for yin
Property sample in be judged as negative ratio, accuracy refers to that true positives and true negative number account for the ratio of tested number.In public affairs
In formula, TP indicates true positives, and TN indicates that true negative, FP indicate false positive, and FN indicates false negative.
In one embodiment, the sample statistics analysis module selected part from single abnormal index according to ROC curve
Or whole abnormal indexes is as united abnormal index.ROC curve refers to Receiver operating curve (receiver
Operating characteristic curve), it is the overall target for reflecting sensibility and specificity continuous variable, is to use structure
Figure method disclose sensibility and specificity correlation, it by the way that continuous variable to be set out to multiple and different critical values, thus
It calculates a series of sensibility and specificities, then by ordinate, (1- specificity) of sensibility be that abscissa is depicted as curve, song
Area is bigger under line, and diagnostic accuracy is higher.On ROC curve, near the upper left point of coordinate diagram for sensibility and specifically
The higher critical value of property.
Although it is worth noting that, the presence of each tumor markers has certain reference value diagnosis is upper,
Also there is the limitation of itself, clinically can make a definite diagnosis liver cancer there is no any single tumor markers at present.And the application proposes
AFP, abnormal prothrombin, carcinomebryonic antigen, carbohydrate antigen 199, carbohydrate antigen 242, multiple tumor markers such as sugar antigen 211
Joint-detection, improves the accuracy of diagnosing cancer of liver, computer-aided diagnosis means can help clinician reduce fail to pinpoint a disease in diagnosis and
Mistaken diagnosis has high clinical value to the discovery of early liver cancer.
Sorter model training module 14 is used to carry out model instruction to multiple sorter models based on the training set data
Practice.
In one embodiment, it based on pretreated data are done through data preprocessing module, to Random Forest model, supports
The sorter models such as vector machine model, BP neural network model, extreme learning machine model carry out model training.Wherein, random gloomy
Woods model is based on Bootstrap method and carries out resampling；The realization function model of algorithm of support vector machine is C-SVC；BP nerve
Network improves network using genetic algorithm, and the input layer of neural network is 5-10 neuron, hidden layer 15-20
A neuron, output layer are 1 neuron.
In one embodiment, different desired outputs is set according to different classes of sample data.With liver cancer group sample
For data, hepatopathy group sample data and health group sample data, the desired output of health group is 0.1, the expectation of hepatopathy group
Output valve is 0.5, and the desired output of liver cancer group is 0.9, and trained model is obtained after successive ignition.
Significantly, since effect of the different classifications device under different classifications scene is different, such as random gloomy
Woods RF is most strong on average, but has also only taken first on 9.9% data set；The average level of support vector machines is tight
With thereafter, first is taken on 10.7% data set.Therefore, it is swollen to provide the liver cancer based on multiple neural networks classifier by the application
Tumor markers aided diagnosis method is examined in accuracy better than the hepatic carcinoma marker auxiliary based on single neural network model
It is disconnected.
Sorter model assessment is used for based on the test set data with test module 15 to the sorter model after training
It is tested, and compared with test result is done superiority and inferiority with the analysis and assessment result of the sample statistics analysis module, is judged accordingly
The validity of tested sorter model.
Specifically, needing to be verified, each trained sorter model when accurate in test set data
When degree index is greater than the analysis and assessment result exported through the sample statistics analysis module, then it is assumed that the sorter model is effective,
Otherwise adjusting parameter, again through the sorter model assessment sorter model new with the training of test module 15.
Classifier diagnostic method application module 16, which is used to utilize, is judged as effective sorter model to tumor markers
Data are detected, and export diagnostic result information accordingly.Such as: patient or doctor can be by the hepatic carcinoma marks of clinical detection
Input data of the will object data as sorter model provided by the present application, provides diagnostic result information by sorter model automatically
It is for reference.
It should be understood that the division of the modules of system above is only a kind of division of logic function, it in actual implementation can be with
Completely or partially it is integrated on a physical entity, it can also be physically separate.And these modules can all be passed through with software
The form that processing element calls is realized；It can also all realize in the form of hardware；Processing element can be passed through with part of module
The form of software is called to realize that part of module passes through formal implementation of hardware.For example, sorter model training module can be single
The processing element solely set up also can integrate and realize in some chip of above system, in addition it is also possible to program code
Form be stored in the memory of above system, called by some processing element of above system and execute the above classifier
The function of model training module.The realization of other modules is similar therewith.Furthermore these modules completely or partially can integrate one
It rises, can also independently realize.Processing element described here can be a kind of integrated circuit, the processing capacity with signal.?
During realization, each step of the above method or the above modules can pass through the integration logic of the hardware in processor elements
The instruction of circuit or software form is completed.
For example, the above module can be arranged to implement one or more integrated circuits of above method, such as:
One or more specific integrated circuits (Application Specific Integrated Circuit, abbreviation ASIC), or,
One or more microprocessors (digital signal processor, abbreviation DSP), or, one or more scene can compile
Journey gate array (Field Programmable Gate Array, abbreviation FPGA) etc..For another example, when some above module passes through place
When managing the form realization of element scheduler program code, which can be general processor, such as central processing unit
(Central Processing Unit, abbreviation CPU) or it is other can be with the processor of caller code.For another example, these modules
It can integrate together, realized in the form of system on chip (system-on-a-chip, abbreviation SOC).
In order to facilitate the understanding of those skilled in the art, hereafter with certain hospital, liver and gall surgical department in March, 2017 in October, 2018
Between more parts of cases as research object.
Sample collection module acquires 139 parts of primary carcinoma of liver cases and 100 parts of non-primary carcinoma of liver cases (including 20
Part benign tumour case and 80 parts of other liver diseases cases), wherein 120 cases are chosen as training set (wherein primary
70 parts of liver cancer case, 50 parts of non-primary carcinoma of liver case), remaining 119 cases are as test set (wherein primary carcinoma of liver case
69 parts, 50 parts of non-primary carcinoma of liver case).
Data preprocessing module rejects the attribute unrelated with patient first, filters out 20 kinds or more abnormal indexes.Using more
The filling that reinforcing method carries out missing values to small part data is refilled, and obtains closing with primary carcinoma of liver using Apriori association algorithm
Data have finally been carried out dimensionality reduction using feature selecting and Principal Component Analysis Algorithm by the abnormal index of strongest 10 kinds or so of connection property
Processing, the hepatic carcinoma marker that finishing screen is selected have alpha-fetoprotein (AFP), carcinomebryonic antigen (CEA), abnormal prothrombin
(PIVKA-II), carbohydrate antigen 199 (CA199), carbohydrate antigen 242 (CA242) and sugar antigen 211 (CA211).
Sample statistics analysis module distinguishes the tumor markers sample data chosen using statistics software spss12.0
Descriptive analysis.Single tumor-marker analyte detection and joint-detection are analyzed respectively.Judged not using non-parametric test
Whether there is statistical significance with the difference between group, is test stone with α=0.05.With diagnostic sensitivity, specificity, accurately
Degree and ROC curve are as evaluation index.
It is analyzed by taking BP neural network as an example, as shown in Fig. 2, showing in the present embodiment to 6 kinds of hepatic carcinoma markers
The ROC curve drawn by research object.Wherein, diagonal line 21 is reference line, and curve 22 indicates sugar antigen 211, curve 23
Indicating that cancer base antigen, curve 24 indicate that carbohydrate antigen 242, curve 25 indicate carbohydrate antigen 199, curve 26 indicates alpha-fetoprotein,
Curve 27 indicates abnormal prothrombin.
Compare the size under each ROC curve it is found that abnormal prothrombin and alpha-fetoprotein ROC curve, i.e., it is bent
Area under line 27 and curve 26 is maximum, so the diagnostic value to liver cancer is maximum；And carbohydrate antigen 242 and sugar antigen 211
ROC curve, i.e. curve 24 and curve 22 is little to the diagnostic value of joint-detection liver cancer close to diagonal line 21, therefore, joint
Detection can abandon the two indexs as training input.Compared with each single index testing result, difference has joint-detection group
Statistical significance (P < 0.05), the following table 1 are the evaluation index tables of sample.
1 sample statistics analytical table of table
Sorter model training module is based on doing pretreated data through data preprocessing module, to random forest mould
The sorter models such as type, supporting vector machine model, BP neural network model, extreme learning machine model carry out model training.Wherein,
Random Forest model is based on Bootstrap method and carries out resampling；The realization function model of algorithm of support vector machine is C-SVC；
BP neural network improves network using genetic algorithm, and the input layer of neural network is 5-10 neuron, hidden layer
For 15-20 neuron, output layer is 1 neuron.
Different desired outputs is set according to different classes of sample data.With liver cancer group sample data, hepatopathy group sample
For notebook data and health group sample data, the desired output of health group is 0.1, and the desired output of hepatopathy group is 0.5, liver
The desired output of cancer group is 0.9, and trained model is obtained after successive ignition.Trained model is needed according to ROC
For the highest tumor markers of Curve selection diagnostic value as joint-detection index, the present embodiment selects alpha-fetoprotein and exception solidifying
Hemase original work are the input pointer of joint-detection.
Sorter model assessment and test module need in test set data each trained sorter model
On verified, when accuracy index be greater than exported through the sample statistics analysis module analysis and assessment result when, then it is assumed that
The sorter model is effective, otherwise adjusting parameter, again through the sorter model assessment classifier new with test module training
Model.Final sorter model is to utilize decision tree thought based on four kinds of classifiers of sorter model training module output
It is built-up.
In this present embodiment, performance of four kinds of classifiers in test set data is as shown in table 2.By table 2 and above
Table 1 compares, and since the accuracy in table 2 has been higher than the accuracy in table 1, therefore thinks that the sorter model is effective.
| Sorter model | Specificity | Sensitivity | Accuracy |
| BP neural network | 98.60% | 76.314% | 89.92% |
| Support vector machines | 98% | 95.65% | 96.64% |
| Random forest RF | 100% | 97.10 | 98.32% |
| Extreme learning machine ELM | 90% | 91.30% | 90.76% |
Classifier diagnostic method application module using be judged as effective sorter model to tumor markers data into
Row detection, exports diagnostic result information accordingly.Such as: patient or doctor can be by the hepatic carcinoma mark numbers of clinical detection
According to the input data as sorter model provided by the present application, diagnostic result information is provided automatically by sorter model for ginseng
It examines.
As shown in figure 3, showing the tumor-marker analyte detection side based on multiple neural networks classifier in one embodiment of the application
The flow diagram of method.
In some embodiments, the method can be applied to controller, such as: ARM controller, FPGA controller, SoC
Controller, dsp controller or MCU controller etc..In some embodiments, the method can also be applied to include depositing
Reservoir, storage control, one or more processing units (CPU), Peripheral Interface, RF circuit, voicefrequency circuit, loudspeaker, Mike
Wind, input/output (I/O) subsystem, display screen, other outputs or the computer for controlling the components such as equipment and outside port；
The computer includes but is not limited to such as desktop computer, laptop, tablet computer, smart phone, smart television, a number
The PCs such as word assistant (Personal Digital Assistant, abbreviation PDA).In other embodiments, described
Method applies also for server, and the server can be arranged in one or more real according to many factors such as function, loads
On body server, it can also be made of server cluster be distributed or concentration.
In this present embodiment, the method includes the steps S31, step S32, step S33, step S34, step S35, step
S36。
In step S31, the sample data of tumor-marker analyte detection sample is acquired, wherein the sample data includes test
Collect data and training set data.
In step s 32, multiple abnormal indexes with tumour with high relevance are filtered out according to the sample data.
In step S33, testing result analysis and assessment are done to single abnormal index based on the sample data data, and/
Or testing result analysis and assessment are done to multiple united abnormal indexes, to generate corresponding analysis and assessment result.
In step S34, model training is carried out to multiple sorter models based on the training set data.
In step s 35, the sorter model after training is tested based on the test set data, and test is tied
Compared with fruit is cooked superiority and inferiority with the analysis and assessment result, the validity of tested sorter model is judged accordingly.
In step S36, tumor markers data are detected using effective sorter model is judged as, accordingly
Export diagnostic result information.
It should be noted that the implementation of the tumor-marker object detecting method based on multiple neural networks classifier of the present embodiment
Mode, it is similar with the embodiment of tumor markers detection system above based on multiple neural networks classifier, thus it is no longer superfluous
It states.
In addition, those of ordinary skill in the art will appreciate that: realize all or part of the steps of above-mentioned each method embodiment
It can be completed by the relevant hardware of computer program.Computer program above-mentioned can store in a computer-readable storage
In medium.When being executed, execution includes the steps that above-mentioned each method embodiment to the program；And storage medium above-mentioned includes:
The various media that can store program code such as ROM, RAM, magnetic or disk.
As shown in figure 4, showing the structural schematic diagram for detecting terminal in another embodiment of the application.The detection that this example provides
Terminal, comprising: processor 41, memory 42, transceiver 43, communication interface 44 and system bus 45；Memory 42 and communication connect
Mouth 44 connect with processor 41 and transceiver 43 and completes mutual communication by system bus 45, and memory 42 is for storing
Computer program, communication interface 44 and transceiver 43 are used for and other equipment are communicated, and processor 41 is for running computer
Program makes to detect each step that terminal executes detection method as above.
System bus mentioned above can be Peripheral Component Interconnect standard (Peripheral Component
Interconnect, abbreviation PCI) bus or expanding the industrial standard structure (Extended Industry Standard
Architecture, abbreviation EISA) bus etc..The system bus can be divided into address bus, data/address bus, control bus etc..
Only to be indicated with a thick line in figure, it is not intended that an only bus or a type of bus convenient for indicating.Communication connects
Mouth is for realizing the communication between database access device and other equipment (such as client, read-write library and read-only library).Storage
Device may include random access memory (Random Access Memory, abbreviation RAM), it is also possible to further include non-volatile deposit
Reservoir (non-volatile memory), for example, at least a magnetic disk storage.
Above-mentioned processor can be general processor, including central processing unit (Central Processing Unit,
Abbreviation CPU), network processing unit (Network Processor, abbreviation NP) etc.；It can also be digital signal processor
(Digital Signal Processing, abbreviation DSP), specific integrated circuit (Application Specific
Integrated Circuit, abbreviation ASIC), field programmable gate array (Field-Programmable Gate Array,
Abbreviation FPGA) either other programmable logic device, discrete gate or transistor logic, discrete hardware components.
In conclusion the application provides tumor markers detection system, method, end based on multiple neural networks classifier
End, medium, the application provides the hepatic carcinoma marker detection scheme based on multiple neural networks classifier, using principal component analysis
With the data preprocessing method of feature extraction, it is based on random forest, support vector machines, BP neural network and extreme learning machine
The assembled classifier model of sorter model finally obtains an accuracy to data training, and specificity, sensitivity meets clinical
The hepatic carcinoma marker classifier of diagnosis helps clinician to reduce misdiagnosis rate in liver cancer first visit, improves hepatic carcinoma mark
Will analyte detection accuracy.So the application effectively overcomes various shortcoming in the prior art and has high industrial utilization value.
The principles and effects of the application are only illustrated in above-described embodiment, not for limitation the application.It is any ripe
Know the personage of this technology all can without prejudice to spirit herein and under the scope of, carry out modifications and changes to above-described embodiment.Cause
This, those of ordinary skill in the art is complete without departing from spirit disclosed herein and institute under technical idea such as
At all equivalent modifications or change, should be covered by claims hereof.
Claims (18)
1. a kind of tumor markers detection system based on multiple neural networks classifier characterized by comprising
Sample collection module, for acquiring the sample data of tumor-marker analyte detection sample, wherein the sample data includes surveying
Examination collection data and training set data；
Data preprocessing module multiple there is the exception of high relevance to refer to tumour for being filtered out according to the sample data
Mark；
Sample statistics analysis module, for doing testing result assessment point to single abnormal index based on the sample data data
Analysis, and/or testing result analysis and assessment are done to multiple united abnormal indexes；
Sorter model training module, for carrying out model training to multiple sorter models based on the training set data；
Sorter model assessment and test module, for being surveyed based on the test set data to the sorter model after training
Examination, and compared with test result is done superiority and inferiority with the analysis and assessment result of the sample statistics analysis module, judgement is tested accordingly
Sorter model validity；
Classifier diagnostic method application module, for using be judged as effective sorter model to tumor markers data into
Row detection, exports diagnostic result information accordingly.
2. system according to claim 1, which is characterized in that the clinical true sample number of sample collection module acquisition
Classify according to and to sample data collected.
3. system according to claim 2, which is characterized in that the tumor markers include hepatic carcinoma marker；Its
In, the mode classification of the sample data includes: that sample data is divided into liver cancer group sample data and non-liver cancer group sample data,
And/or it is divided into liver cancer group sample data, hepatopathy group sample data and health group sample data.
4. system according to claim 1, which is characterized in that the data preprocessing module is used to be based on principal component analysis
Algorithm carries out feature selecting, to carry out dimension-reduction treatment to the multiple abnormal indexes gone out according to the sample data initial screening, and
Obtain the abnormal index with tumour with high relevance.
5. system according to claim 4, which is characterized in that described to include with tumour has high relevance abnormal index
With hepatic carcinoma have high relevance abnormal index comprising alpha-fetoprotein, further includes: abnormal prothrombin, carcinomebryonic antigen,
Carbohydrate antigen 199, carbohydrate antigen 242, sugar antigen 211, carbohydrate antigen 125, sialic acid or ferritin.
6. system according to claim 1, which is characterized in that the evaluation index of testing result analysis and assessment includes sensitive
Degree, specificity and accuracy；Wherein, sensitivity indicates the practical ratio to be judged as the positive in positive sample, specificity table
Show the practical ratio to be judged as feminine gender in negative sample, accuracy indicates that true positives and true negative number account for tested total number of persons
Ratio.
7. system according to claim 1, which is characterized in that the sorter model includes: Random Forest model, supports
Vector machine model, BP neural network model, extreme learning machine model.
8. system according to claim 1 characterized by comprising the sample statistics analysis module is according to ROC curve
Selected part or the abnormal index of whole are as united abnormal index from single abnormal index.
9. a kind of tumor-marker object detecting method based on multiple neural networks classifier characterized by comprising
The sample data for acquiring tumor-marker analyte detection sample, wherein the sample data includes test set data and training set number
According to；
Multiple abnormal indexes with tumour with high relevance are filtered out according to the sample data；
Testing result analysis and assessment are done to single abnormal index based on the sample data data, and/or to multiple united different
Chang Zhibiao does testing result analysis and assessment, to generate corresponding analysis and assessment result；
Model training is carried out to multiple sorter models based on the training set data；
The sorter model after training is tested based on the test set data, and by test result and the analysis and assessment
As a result it does superiority and inferiority to compare, judges the validity of tested sorter model accordingly；
Tumor markers data are detected using effective sorter model is judged as, export diagnostic result letter accordingly
Breath.
10. according to the method described in claim 9, it is characterized in that, which comprises the clinical true sample data of acquisition
And classify to sample data collected.
11. according to the method described in claim 10, it is characterized in that, the tumor markers include hepatic carcinoma marker；
Wherein, the mode classification of the sample data includes: that sample data is divided into liver cancer group sample data and non-liver cancer group sample number
According to, and/or it is divided into liver cancer group sample data, hepatopathy group sample data and health group sample data.
12. according to the method described in claim 9, it is characterized in that, which comprises based on Principal Component Analysis Algorithm carry out
Feature selecting, to carry out dimension-reduction treatment to the multiple abnormal indexes gone out according to the sample data initial screening, and described in acquisition
There is the abnormal index of high relevance with tumour.
13. according to the method for claim 12, which is characterized in that the abnormal index packet with tumour with high relevance
Include the abnormal index that there is high relevance with hepatic carcinoma comprising alpha-fetoprotein, further includes: abnormal prothrombin, cancer embryo are anti-
Original, carbohydrate antigen 199, carbohydrate antigen 242, sugar antigen 211, carbohydrate antigen 125, sialic acid or ferritin.
14. according to the method described in claim 9, it is characterized in that, the evaluation index of testing result analysis and assessment includes sensitive
Degree, specificity and accuracy；Wherein, sensitivity indicates the practical ratio to be judged as the positive in positive sample, specificity table
Show the practical ratio to be judged as feminine gender in negative sample, accuracy indicates that true positives and true negative number account for tested total number of persons
Ratio.
15. according to the method described in claim 9, it is characterized in that, the sorter model includes: Random Forest model, branch
Hold vector machine model, BP neural network model, extreme learning machine model.
16. according to the method described in claim 9, it is characterized in that, which comprises according to ROC curve from single exception
Selected part or the abnormal index of whole are as united abnormal index in index.
17. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The tumor-marker quality testing based on multiple neural networks classifier described in any one of claim 9 to 16 is realized when being executed by processor
Survey method.
18. a kind of detection terminal characterized by comprising processor and memory；
The memory is for storing computer program；
The processor is used to execute the computer program of the memory storage, so that the terminal executes such as claim 9
To the tumor-marker object detecting method described in any one of 16 based on multiple neural networks classifier.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910149298.5A CN110031624A (en) | 2019-02-28 | 2019-02-28 | Tumor markers detection system based on multiple neural networks classifier, method, terminal, medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910149298.5A CN110031624A (en) | 2019-02-28 | 2019-02-28 | Tumor markers detection system based on multiple neural networks classifier, method, terminal, medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110031624A true CN110031624A (en) | 2019-07-19 |
Family
ID=67235717
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910149298.5A Pending CN110031624A (en) | 2019-02-28 | 2019-02-28 | Tumor markers detection system based on multiple neural networks classifier, method, terminal, medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110031624A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110991536A (en) * | 2019-12-02 | 2020-04-10 | 上海应用技术大学 | Training method of early warning model of primary liver cancer |
| CN111326260A (en) * | 2020-01-09 | 2020-06-23 | 上海中科新生命生物科技有限公司 | Medical analysis method, device, equipment and storage medium |
| CN111413498A (en) * | 2020-04-08 | 2020-07-14 | 复旦大学附属中山医院 | Autoantibody 7-AAb detection panel for hepatocellular carcinoma and application thereof |
| CN111524594A (en) * | 2020-06-12 | 2020-08-11 | 山东大学 | Target population blood system malignant tumor screening system |
| CN111797786A (en) * | 2020-07-09 | 2020-10-20 | 郑州中普医疗器械有限公司 | Detection method for in vitro biological samples, four-classification, computer device and computer-readable storage medium |
| CN111863248A (en) * | 2020-08-04 | 2020-10-30 | 桂林电子科技大学 | Effective method for constructing clinical decision model |
| CN112862018A (en) * | 2021-04-19 | 2021-05-28 | 温州医科大学 | Tumor classification device based on 5hmC modified lncRNA |
| WO2022027714A1 (en) * | 2020-08-06 | 2022-02-10 | 深圳市第二人民医院(深圳市转化医学研究院) | Diagnostic capability evaluation method, system and device for diagnosis model, and storage medium |
| CN114550832A (en) * | 2022-02-24 | 2022-05-27 | 武汉金开瑞生物工程有限公司 | Methods, systems, and media for holistic screening of proteomic clinical biomarkers |
| CN115856313A (en) * | 2022-11-17 | 2023-03-28 | 吉林大学 | Marker for diagnosing thyroid papillary carcinoma and predicting cervical lymph node metastasis |
| CN116312778A (en) * | 2023-01-19 | 2023-06-23 | 广州医科大学 | Auxiliary diagnosis and prediction method, device, equipment and medium for mature B cell tumor |
| CN118186086A (en) * | 2024-03-21 | 2024-06-14 | 山东一点基因科技有限公司 | Comprehensive analysis device for head and neck tumor marker detection data |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107024586A (en) * | 2017-04-20 | 2017-08-08 | 中国人民解放军第五九医院 | Method based on artificial neural network tumor-marker joint-detection auxiliary diagnosis liver cancer |
| CN107480474A (en) * | 2017-08-01 | 2017-12-15 | 山东师范大学 | Grader modeling evaluation method of calibration and system based on gut flora abundance |
| TW201804348A (en) * | 2016-07-29 | 2018-02-01 | 長庚醫療財團法人林口長庚紀念醫院 | Method for analyzing cancer detection result by establishing cancer prediction model and combining tumor marker kits analyzing the cancer detection result by using the established cancer prediction model and combining the detection results of the tumor marker kits |
| CN109345302A (en) * | 2018-09-27 | 2019-02-15 | 腾讯科技（深圳）有限公司 | Machine learning model training method, device, storage medium and computer equipment |
-
2019
- 2019-02-28 CN CN201910149298.5A patent/CN110031624A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW201804348A (en) * | 2016-07-29 | 2018-02-01 | 長庚醫療財團法人林口長庚紀念醫院 | Method for analyzing cancer detection result by establishing cancer prediction model and combining tumor marker kits analyzing the cancer detection result by using the established cancer prediction model and combining the detection results of the tumor marker kits |
| CN107024586A (en) * | 2017-04-20 | 2017-08-08 | 中国人民解放军第五九医院 | Method based on artificial neural network tumor-marker joint-detection auxiliary diagnosis liver cancer |
| CN107480474A (en) * | 2017-08-01 | 2017-12-15 | 山东师范大学 | Grader modeling evaluation method of calibration and system based on gut flora abundance |
| CN109345302A (en) * | 2018-09-27 | 2019-02-15 | 腾讯科技（深圳）有限公司 | Machine learning model training method, device, storage medium and computer equipment |
Non-Patent Citations (2)
| Title |
|---|
| ALI KALANTARI等: "Computational intelligence approaches for classification of medical data: State-of-the-art, future challenges and research directions", 《NEUROCOMPUTING》 * |
| YANWU CHU等: "Discrimination of nasopharyngeal carcinoma serum using laser-induced breakdown spectroscopy combined with an extreme learning machine and random forest method", 《JAAS》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110991536B (en) * | 2019-12-02 | 2023-05-09 | 上海应用技术大学 | Training method of early warning model of primary liver cancer |
| CN110991536A (en) * | 2019-12-02 | 2020-04-10 | 上海应用技术大学 | Training method of early warning model of primary liver cancer |
| CN111326260A (en) * | 2020-01-09 | 2020-06-23 | 上海中科新生命生物科技有限公司 | Medical analysis method, device, equipment and storage medium |
| CN111413498A (en) * | 2020-04-08 | 2020-07-14 | 复旦大学附属中山医院 | Autoantibody 7-AAb detection panel for hepatocellular carcinoma and application thereof |
| CN111413498B (en) * | 2020-04-08 | 2023-08-04 | 复旦大学附属中山医院 | Autoantibody 7-AAb detection panel for liver cell liver cancer and application thereof |
| CN111524594A (en) * | 2020-06-12 | 2020-08-11 | 山东大学 | Target population blood system malignant tumor screening system |
| CN111797786A (en) * | 2020-07-09 | 2020-10-20 | 郑州中普医疗器械有限公司 | Detection method for in vitro biological samples, four-classification, computer device and computer-readable storage medium |
| CN111863248B (en) * | 2020-08-04 | 2022-10-11 | 桂林电子科技大学 | Effective method for constructing clinical decision model |
| CN111863248A (en) * | 2020-08-04 | 2020-10-30 | 桂林电子科技大学 | Effective method for constructing clinical decision model |
| WO2022027714A1 (en) * | 2020-08-06 | 2022-02-10 | 深圳市第二人民医院(深圳市转化医学研究院) | Diagnostic capability evaluation method, system and device for diagnosis model, and storage medium |
| CN112862018B (en) * | 2021-04-19 | 2022-09-02 | 温州医科大学 | Tumor classification device based on 5hmC modified lncRNA |
| CN112862018A (en) * | 2021-04-19 | 2021-05-28 | 温州医科大学 | Tumor classification device based on 5hmC modified lncRNA |
| CN114550832A (en) * | 2022-02-24 | 2022-05-27 | 武汉金开瑞生物工程有限公司 | Methods, systems, and media for holistic screening of proteomic clinical biomarkers |
| CN115856313A (en) * | 2022-11-17 | 2023-03-28 | 吉林大学 | Marker for diagnosing thyroid papillary carcinoma and predicting cervical lymph node metastasis |
| CN116312778A (en) * | 2023-01-19 | 2023-06-23 | 广州医科大学 | Auxiliary diagnosis and prediction method, device, equipment and medium for mature B cell tumor |
| CN116312778B (en) * | 2023-01-19 | 2024-02-13 | 广州医科大学 | Auxiliary diagnosis and prediction method, device, equipment and medium for mature B cell tumor |
| CN118186086A (en) * | 2024-03-21 | 2024-06-14 | 山东一点基因科技有限公司 | Comprehensive analysis device for head and neck tumor marker detection data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110031624A (en) | Tumor markers detection system based on multiple neural networks classifier, method, terminal, medium | |
| Wang et al. | Risk assessment of coronary heart disease based on cloud-random forest | |
| Zhou et al. | In-depth mining of clinical data: the construction of clinical prediction model with R | |
| Alksas et al. | A novel computer-aided diagnostic system for accurate detection and grading of liver tumors | |
| Ley et al. | Idiopathic pulmonary fibrosis: CT and risk of death | |
| Chetty et al. | Role of attributes selection in classification of Chronic Kidney Disease patients | |
| CN114926477B (en) | Brain tumor multi-mode MRI image segmentation method based on deep learning | |
| Luo et al. | DiagTest3Grp: an R package for analyzing diagnostic tests with three ordinal groups | |
| CN110444248A (en) | Cancer Biology molecular marker screening technique and system based on network topology parameters | |
| CN112633601B (en) | Method, device, equipment and computer medium for predicting disease event occurrence probability | |
| CN106248945B (en) | Method, system and the kit of the grouping of hepatocellular carcinoma radical excision prognosis situation are carried out to patients with hepatocellular carcinoma | |
| CN115099331A (en) | Auxiliary diagnosis system for malignant pleural effusion based on interpretable machine learning algorithm | |
| Alsheh Ali et al. | Detection of potential microcalcification clusters using multivendor for‐presentation digital mammograms for short‐term breast cancer risk estimation | |
| CN113128654A (en) | Improved random forest model for coronary heart disease pre-diagnosis and pre-diagnosis system thereof | |
| Zhao et al. | Current progress in artificial intelligence-assisted medical image analysis for chronic kidney disease: A literature review | |
| CN107169264B (en) | complex disease diagnosis system | |
| Nashwan et al. | Using artificial intelligence to improve body iron quantification: A scoping review | |
| Liang et al. | HTC-Net: Hashimoto’s thyroiditis ultrasound image classification model based on residual network reinforced by channel attention mechanism | |
| Cui et al. | Value of breast MRI omics features and clinical characteristics in Breast Imaging Reporting and Data System (BI-RADS) category 4 breast lesions: an analysis of radiomics-based diagnosis | |
| Wang et al. | Survival risk prediction model for ESCC based on relief feature selection and CNN | |
| Fanjul-Hevia et al. | A non-parametric test for comparing conditional ROC curves | |
| CN116738352B (en) | Method and device for classifying abnormal rod cells of retinal vascular occlusion disease | |
| Scicolone et al. | Radiomics and artificial intelligence: General notions and applications in the carotid vulnerable plaque | |
| CN110163195A (en) | Liver cancer divides group's prediction model, its forecasting system and liver cancer to divide group's judgment method | |
| Seetharam et al. | Broadening Perspectives of Artificial Intelligence in Echocardiography |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190719 |
|
| WD01 | Invention patent application deemed withdrawn after publication |