CN107526946A - Merge the gene expression data cancer classification method of self study and low-rank representation - Google Patents
Merge the gene expression data cancer classification method of self study and low-rank representation Download PDFInfo
- Publication number
- CN107526946A CN107526946A CN201611207518.8A CN201611207518A CN107526946A CN 107526946 A CN107526946 A CN 107526946A CN 201611207518 A CN201611207518 A CN 201611207518A CN 107526946 A CN107526946 A CN 107526946A
- Authority
- CN
- China
- Prior art keywords
- mrow
- matrix
- low
- rank
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
- G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
Landscapes
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Medical Informatics (AREA)
- Biophysics (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Epidemiology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Public Health (AREA)
- Bioethics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Biotechnology (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Theoretical Computer Science (AREA)
- Investigating Or Analysing Biological Materials (AREA)
- Measuring Or Testing Involving Enzymes Or Micro-Organisms (AREA)
Abstract
The invention discloses a kind of gene expression data cancer classification method for merging self study and low-rank representation, including:Step 1, data set is expressed for given cancer gene, data are merged into structure data matrix, and make normalized;Step 2, the data matrix for obtaining, are decomposed using low-rank expression, obtain a low-rank matrix and a sparse matrix;Step 3, the label information using training set, calculate the initial point of each classification respectively in low-rank matrix and sparse matrix;Step 4, a kind of unsupervised clustering is used in low-rank matrix and sparse matrix respectively, obtain the prediction result based on low-rank matrix and sparse matrix respectively;Two step 5, contrast prediction results, if without prediction identical sample or reaching maximum iteration, export the prediction result based on low-rank expression matrix;Otherwise, prediction identical sample is removed into test set and adds training set, return to step 3.Precision of prediction can be improved in the case of using a small amount of mark sample using the present invention, reduce the time in mark sample and human cost.
Description
Technical field
The present invention relates to bioinformatics gene expression and cancer classification field, is a kind of fusion self study specifically
With the gene expression data cancer classification method of low-rank representation.
Background technology
Cancer is a kind of caused fatal disease due to cell abnormal growth, so far, is not had yet fully effective
Treatment method.Early diagnosis can must effectively help treatment of cancer, so accurately classification prediction how is carried out to cancer is
One has the problem of researching value very much.With the development of high-throughput techniques, the gene expression data about cancer is promptly
Accumulation, while machine learning techniques also obtain significant progress in recent years, therefore utilize gene expression data and machine learning
To predict that cancer class is possibly realized, such as:(1)Chen,X.Y.and Jian,C.R.Gene expression data
clustering based on graph regularized subspace segmentation.Neurocomputing
2014;143:44-50.(2)Liao,Q.,Guan,N.and Zhang,Q.Gauss-Seidel based non-negative
matrix factorization for gene expression clustering.In,2016IEEE International
Conference on Acoustics,Speech and Signal Processing(ICASSP).IEEE;
2016.p.2364-2368.(3)Liu,J.X.,et al.RPCA-Based Tumor Classification Using Gene
Expression Data.IEEE ACM T Comput Bi 2015;12(4):964-970. wait.However, existing method is big
Mostly unsupervised approaches and measure of supervision, both suffer from respective defect.
Unsupervised approaches learning method finds potential structure by proposing a model from without label data.By institute
Some samples all do not mark, so can not carry out error correction using label information in the training of model.It is unsupervised
This characteristic of study causes the predictive ability of model weaker, can not provide effective precision of prediction.Supervised learning method and nothing
Supervised learning on the contrary, it by using there is label data to carry out training pattern.Due to label data can be used in training, supervision
The model that learning method obtains can provide higher precision of prediction.But a model is trained to need to use greatly using measure of supervision
The labeled data of amount, and the cost of labeled data is often sufficiently expensive and needs to expend substantial amounts of manpower and time, it is especially right
Gene expression data is labeled.
In view of the defects of two kinds of learning methods suffer from overcoming, the proposition of semi-supervised learning carries to solve the above problems
A new approaches are supplied:Using largely without label data and on a small quantity thering is label data to be provided come training pattern, the model
It is much better than the prediction effect of unsupervised approaches.It is a kind of traditional semi-supervised learning side that the improved self learning model of this method, which is exactly,
Method, it by the higher sample of confidence level in prediction by adding training set, the training and prediction of continuous iteration, finally to all surveys
The data that examination is concentrated are classified.Nowadays, existing many effective semi-supervised methods are used for point of cancer gene expression data
Analysis, such as:(1)Cai,X.F.,et al.Local and Global Preserving Semi-supervised
Dimensionality Reduction Based on Random Subspace for Cancer
Classification.IEEE J Biomed Health 2014;18(2):500-507.(2)Halder,A.and Misra,
S.Semi-supervised fuzzy K-NN for cancer classification from microarray gene
expression data.2014First International Conference on Automation,Control,
Energy&Systems(ACES-14)2014:266-270. wait.However, the processing of gene expression data still leaves challenge:
(1) gene expression data has very high dimension
Because each feature of gene expression data corresponds to a gene, and the mankind have no less than 2.5 ten thousand genes, institute
Often there are tens thousand of individual characteristic components with gene expression data.Traditional sorting technique is when handling high dimensional data, in data
Noise and redundancy are very sensitive, it is difficult to provide accurate prediction;
(2) the data set very little of gene expression data
Because using the gene expression of gene microarray technology measure, costly, time and human cost are very high, therefore, one
The data set very little that secondary property obtains, usually only includes tens or hundreds of samples, and too small data volume is difficult to train effectively
Model.
The content of the invention
It is an object of the invention to provide a kind of gene expression data cancer classification side for merging self study and low-rank representation
Method, solve the problems, such as that carrying out cancer classification prediction using gene expression data in the prior art is present:Data dimension is high, test set
Small and labeled data is few.
The technical solution for realizing the object of the invention is:A kind of gene expression data for merging self study and low-rank representation
Cancer classification method, comprises the following steps:
Step 1, data set is expressed for given cancer gene, wherein the collection for having label data is combined into training set, no label
Data acquisition system is test set;Data are merged into structure data matrix X, and make normalized;
Step 2, the data matrix for obtaining, are decomposed using low-rank expression, obtain a low-rank matrix Z and
One sparse matrix E;
Step 3, the label information using training set, calculate each classification i respectively on low-rank matrix Z and sparse matrix E
Initial point coordinates p(i);
Step 4, a kind of unsupervised clustering is used on low-rank matrix Z and sparse matrix E respectively, be based on respectively
Low-rank matrix Z and sparse matrix E prediction result lZAnd lE;
Two step 5, contrast prediction result lZAnd lEIf without prediction identical sample or reaching maximum iteration, export
Prediction result l based on low-rank expression matrixZ;Otherwise, prediction identical sample is removed into test set and adds training set, returned
Step 3.
Compared with prior art, its remarkable advantage is the present invention:1) low-rank representation method, Ke Yicong are combined in this method
The global characteristics of essence are extracted in original high dimensional data;(2) this method used simultaneously decompose to obtain in low-rank representation it is low
Order matrix information and sparse matrix information, than traditional method (merely with the information in a matrix) based on low-rank representation more
To be effective.
Brief description of the drawings
Fig. 1 is an exemplary flow of the gene expression data cancer classification method of fusion self study and low-rank representation
Figure.
Fig. 2 be a certain cancer gene express data set on schematic diagram, (a), (b), (c) be respectively former data matrix and
Low-rank matrix, sparse matrix after low-rank decomposition, the gray value of each corresponding pixel of expression value in matrix.On each matrix
In the horizontal stripe of side, the classification of the corresponding cancer of each color lump.
Embodiment
It is a typical high dimensional and small sample size problem to carry out cancer classification prediction using gene expression data.In order to solve this
Individual problem, low-rank representation is used for reference, a kind of feature extracting method for being usually used in matrix recovery in image processing field, it passes through about
Beam data rank of matrix, to obtain the essential low dimensional structures of data.
By using semi-supervised learning method and feature extracting method, can solve to carry out cancer using gene expression data
The problem of disease classification prediction.
Embodiments of the present invention are described in detail by way of example below in conjunction with accompanying drawing.
As shown in figure 1, according to the preferred embodiment of the present invention, the gene expression data cancer of fusion self study and low-rank representation
Disease sorting technique, the sample for expressing a cancer gene in data set carry out class prediction.In order to reflect practical application
Scene, regard which part data as no label data, and be defined as test set;Remaining sample set is defined as training
Collection.During training and prediction, the label information of sample in training set can only be used, the classification information of test set is used for and surveyed
The prediction classification of examination collection compares..Classification prediction is divided into two stages:Feature extraction phases and training and forecast period, below
With reference to shown in Fig. 1, the realization in above-mentioned two stage is described in detail.
(1) feature extraction phases
The first step, by training set and test set characteristic vector merge, build a data matrix X, the data of acquisition
Matrix X needs to meet:
X=[x1,x2,…,xn]∈Rd×n
Wherein, xiFor the column vector of a gene expression data sample, vector dimension d.N sample is shared in X, n is
The sum of sample number in training set and test set.Each vector is needed by normalized.
Second step, for the given data matrix X, decomposed using low-rank expression, obtained low-rank square
Battle array Z and sparse matrix E need to meet following condition:
S.t., X=XZ+E
Wherein, Z=[z1,z2,...,zn]∈Rn×n,||Z||*=∑iσi(Z) nuclear norm for being Z, σi(Z) it is i-th of Z
Singular value;Refer to E l2,1Norm;λ is balance parameters.In this example, select λ=
2 can obtain preferable effect.Alternating direction Multiplier Algorithm (Alternating Direction can be used by solving above formula
Method of Multipliers)。
Fig. 2 is to express the schematic diagram on data set in a certain cancer gene, and the data set includes 14 kinds of cancer types altogether
(BR:Breast cancer, PR:Prostate cancer, LU:Lung cancer, CO:Colon and rectum gland cancer, LY:Lymph cancer, BL:TCCB, ML:
Melanoma, UT:Adenocarcinoma of endometrium, LE:Leukaemia, RE:Clear-cell carcinoma, PA:Cancer of pancreas, OV:Adenocarcinoma ovaries, MS:Pleura
Celiothelioma, CNS:Central nervous system cancer), 198 samples, each sample shares 11370 characteristic components.As can be seen that
(a) outstanding feature is not present in data distribution in;(b) it is obvious that the cancer sample of identical category falls the characteristics of data distribution in
Enter in same sub-spaces, the classification based on the matrix obviously can be better than the classification based on original matrix;(c) data distribution in
There is the characteristics of certain, due to openness, nonzero value is seldom in matrix, and those have the characteristic component of more nonzero value can be to final
The result of classification causes bigger influence.Classification based on sparse matrix can provide a different visual angle, to final classification
As a result booster action is played.
(2) training and forecast period
Different from the characteristics of training and prediction separate in traditional measure of supervision, this method is a kind of semi-supervised clustering side
Method, the training and prediction of model are carried out simultaneously.As shown in figure 1, training and prediction iteration are carried out, three phases are divided into.
The first step, initial point is tried to achieve respectively to matrix Z and matrix E.Each classification i initial point coordinates in ZCalculating side
Formula is as follows:
Wherein,For ZlIn points in i-th of cluster,For ZlIn i-th of cluster j-th of sample, ZlTo have label in Z
The matrix of sample composition;Can be in the hope of the initial point coordinates of each classification i in E with same method
Second step, based on the initial point tried to achieve in the first stepWithNothing is used to low-rank matrix Z and sparse matrix E
Supervised Clustering Methods.In this example, the K-means algorithms of selection standard are as unsupervised clustering, distance metric choosing
Preferable result can be obtained by selecting Minkowski distances.Based on low-rank matrix and the prediction result of sparse matrix vector difference
It is expressed as lZAnd lE。
Two 3rd step, contrast prediction result lZAnd lE, the suitable sample that do not mark of selection is as under mark sample and addition
An iteration;Or terminate iteration, output result.
Select the mode of suitable sample as follows:
(1) one does not mark sample and is chosen as marking sample addition next iteration, and and if only if, and following formula is set up:
Wherein,It is cluster result lZIn i-th of sample prediction result,It is cluster result lEIn the i-th sample prediction
As a result;
(2) a set S is defined, empty set is initialized with, all samples for meeting above-mentioned standard is put into wherein.
The specific standards whether evaluation algorithm terminates are as follows:
(1) if iterations reaches maximum times, termination algorithm;Otherwise, prediction identical sample is removed into test set simultaneously
Training set is added, into next iteration;
(2) if S is empty set, termination algorithm;Otherwise, prediction identical sample is removed into test set and adds training set, entered
Enter next iteration.
If algorithm does not terminate updates training set and test set as follows:If sample i, will in SFrom ZuMiddle removal,
And add ZlIn;WillFrom EuMiddle removal, and add ElIn.Wherein, ZlFor the matrix being made up of in Z exemplar, ZuFor nothing in Z
The matrix of exemplar composition;ElFor the matrix being made up of in E exemplar, EuThe matrix formed for unlabeled exemplars in E.
After algorithm terminates, l is returnedZExported as prediction result.
Claims (9)
- A kind of 1. gene expression data cancer classification method for merging self study and low-rank representation, it is characterised in that including following step Suddenly:Step 1, data set is expressed for given cancer gene, wherein the collection for having label data is combined into training set, no label data Collection is combined into test set;Data are merged into structure data matrix X, and make normalized;Step 2, the data matrix for obtaining, are decomposed using low-rank expression, obtain a low-rank matrix Z and one Sparse matrix E;Step 3, the label information using training set, calculate the first of each classification i respectively on low-rank matrix Z and sparse matrix E Initial point coordinate p(i);Step 4, a kind of unsupervised clustering is used on low-rank matrix Z and sparse matrix E respectively, obtain be based on low-rank respectively Matrix Z and sparse matrix E prediction result lZAnd lE;Two step 5, contrast prediction result lZAnd lEIf without prediction identical sample or reaching maximum iteration, output is based on The prediction result l of low-rank expression matrixZ;Otherwise, prediction identical sample is removed into test set and adds training set, return to step 3。
- 2. fusion self study according to claim 1 and the gene expression data cancer classification method of low-rank representation, it is special Sign is:Cancer gene expression data set is given described in step 1 to include label data and be without label data, wherein label Cancer class.
- 3. fusion self study according to claim 1 and the gene expression data cancer classification method of low-rank representation, it is special Sign is, in the step 1, the data matrix X of acquisition needs to meet:X=[x1,x2,…,xn]∈Rd×nWherein, xiFor the column vector of a gene expression data sample, n sample is shared in vector dimension d, X, n is training set With the sum of sample number in test set;Each vector is needed by normalized.
- 4. fusion self study according to claim 1 and the gene expression data cancer classification method of low-rank representation, it is special Sign is:In the step 2, for given data matrix X, decomposed using low-rank expression, obtained low-rank Matrix Z and sparse matrix E need to meet following condition:<mrow> <munder> <mrow> <mi>m</mi> <mi>i</mi> <mi>n</mi> </mrow> <mrow> <mi>Z</mi> <mo>,</mo> <mi>E</mi> </mrow> </munder> <mo>|</mo> <mo>|</mo> <mi>Z</mi> <mo>|</mo> <msub> <mo>|</mo> <mo>*</mo> </msub> <mo>+</mo> <mi>&lambda;</mi> <mo>|</mo> <mo>|</mo> <mi>E</mi> <mo>|</mo> <msub> <mo>|</mo> <mrow> <mn>2</mn> <mo>,</mo> <mn>1</mn> </mrow> </msub> </mrow>S.t., X=XZ+EWherein, | | Z | |*=∑iσi(Z) nuclear norm for being Z, σi(Z) i-th of singular value for being Z;Refer to E l2,1Norm;λ is balance parameters.
- 5. fusion self study according to claim 1 and the gene expression data cancer classification method of low-rank representation, it is special Sign is:In the step 3, each classification i initial point coordinates in ZCalculation is as follows:<mrow> <msubsup> <mi>p</mi> <mi>z</mi> <mrow> <mo>(</mo> <mi>i</mi> <mo>)</mo> </mrow> </msubsup> <mo>=</mo> <mfrac> <mrow> <munderover> <mo>&Sigma;</mo> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <msubsup> <mi>n</mi> <mi>l</mi> <mrow> <mo>(</mo> <mi>i</mi> <mo>)</mo> </mrow> </msubsup> </munderover> <msubsup> <mi>z</mi> <mrow> <mi>l</mi> <mo>,</mo> <mi>j</mi> </mrow> <mrow> <mo>(</mo> <mi>i</mi> <mo>)</mo> </mrow> </msubsup> </mrow> <msubsup> <mi>n</mi> <mi>l</mi> <mrow> <mo>(</mo> <mi>i</mi> <mo>)</mo> </mrow> </msubsup> </mfrac> </mrow>Wherein,For ZlIn points in i-th of cluster,For ZlIn i-th of cluster j-th of sample, ZlTo have exemplar in Z The matrix of composition;Each classification i initial point coordinates in ECalculation is as follows::<mrow> <msubsup> <mi>p</mi> <mi>E</mi> <mrow> <mo>(</mo> <mi>i</mi> <mo>)</mo> </mrow> </msubsup> <mo>=</mo> <mfrac> <mrow> <munderover> <mo>&Sigma;</mo> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <msubsup> <mi>n</mi> <mi>l</mi> <mrow> <mo>(</mo> <mi>i</mi> <mo>)</mo> </mrow> </msubsup> </munderover> <msubsup> <mi>e</mi> <mrow> <mi>l</mi> <mo>,</mo> <mi>j</mi> </mrow> <mrow> <mo>(</mo> <mi>i</mi> <mo>)</mo> </mrow> </msubsup> </mrow> <msubsup> <mi>n</mi> <mi>l</mi> <mrow> <mo>(</mo> <mi>i</mi> <mo>)</mo> </mrow> </msubsup> </mfrac> </mrow>Wherein,For ElIn i-th of cluster j-th of sample, ElFor the matrix being made up of in E exemplar.
- 6. fusion self study according to claim 1 and the gene expression data cancer classification method of low-rank representation, it is special Sign is:In the step 4, a kind of unsupervised clustering is used low-rank matrix Z and sparse matrix E, and this method needs true Fixed initial cluster centre, and select a kind of distance metric to weigh the similarity of two samples, in this step, initially Put and be respectivelyWithDistance metric is Minkowski distances.
- 7. fusion self study according to claim 1 and the gene expression data cancer classification method of low-rank representation, it is special Sign is:In the step 5, two prediction result l are contrastedZAnd lE, select not marking sample as mark sample and add next Secondary iteration, its specific implementation are as follows:(1) one does not mark sample and is chosen as marking sample addition next iteration, and and if only if, and following formula is set up:<mrow> <msub> <mi>l</mi> <msub> <mi>z</mi> <mi>i</mi> </msub> </msub> <mo>=</mo> <msub> <mi>l</mi> <msub> <mi>e</mi> <mi>i</mi> </msub> </msub> </mrow>Wherein,It is cluster result lZIn i-th of sample prediction result,It is cluster result lEIn the i-th sample prediction knot Fruit;(2) a set S is defined, empty set is initialized with, all samples for meeting above-mentioned standard is put into wherein.
- 8. fusion self study according to claim 1 and the gene expression data cancer classification method of low-rank representation, it is special Sign is:In the step 5, the specific standards whether evaluation algorithm terminates are as follows:(1) if iterations reaches maximum times, termination algorithm;Otherwise, prediction identical sample is removed into test set and added Training set, return to step 3, into next iteration;(2) if S is empty set, termination algorithm;Otherwise, prediction identical sample is removed into test set and adds training set, return to step Rapid 3, into next iteration;After algorithm terminates, l is returnedZExported as prediction result.
- 9. the gene expression data cancer classification method of the fusion self study and low-rank representation according to claim 7 or 8, its It is characterised by:In the step 5, prediction identical sample is removed into test set and adds being defined as follows for training set:If sample i, will in SFrom ZuMiddle removal, and add ZlIn;WillFrom EuMiddle removal, and add ElIn;Wherein, ZlFor Z In the matrix that is made up of exemplar, ZuThe matrix formed for unlabeled exemplars in Z;ElFor the square being made up of in E exemplar Battle array, EuThe matrix formed for unlabeled exemplars in E.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611207518.8A CN107526946B (en) | 2016-12-23 | 2016-12-23 | Gene expression data cancer classification method combining self-learning and low-rank representation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611207518.8A CN107526946B (en) | 2016-12-23 | 2016-12-23 | Gene expression data cancer classification method combining self-learning and low-rank representation |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107526946A true CN107526946A (en) | 2017-12-29 |
CN107526946B CN107526946B (en) | 2021-07-06 |
Family
ID=60748589
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611207518.8A Active CN107526946B (en) | 2016-12-23 | 2016-12-23 | Gene expression data cancer classification method combining self-learning and low-rank representation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107526946B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108169728A (en) * | 2018-01-12 | 2018-06-15 | 西安电子科技大学 | Range extension target detection method based on Minkowski distances |
CN109378039A (en) * | 2018-08-20 | 2019-02-22 | 中国矿业大学 | Oncogene based on discrete constraint and the norm that binds expresses spectral-data clustering method |
CN109671468A (en) * | 2018-12-13 | 2019-04-23 | 韶关学院 | A kind of feature gene selection and cancer classification method |
CN109903166A (en) * | 2018-12-25 | 2019-06-18 | 阿里巴巴集团控股有限公司 | A kind of data Risk Forecast Method, device and equipment |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102722892A (en) * | 2012-06-13 | 2012-10-10 | 西安电子科技大学 | SAR (synthetic aperture radar) image change detection method based on low-rank matrix factorization |
CN103400143A (en) * | 2013-07-12 | 2013-11-20 | 中国科学院自动化研究所 | Data subspace clustering method based on multiple view angles |
US20140025689A1 (en) * | 2012-04-24 | 2014-01-23 | International Business Machines Corporation | Determining a similarity between graphs |
CN103793600A (en) * | 2014-01-16 | 2014-05-14 | 西安电子科技大学 | Isolated component analysis and linear discriminant analysis combined cancer forecasting method |
CN105427296A (en) * | 2015-11-11 | 2016-03-23 | 北京航空航天大学 | Ultrasonic image low-rank analysis based thyroid lesion image identification method |
CN106096654A (en) * | 2016-06-13 | 2016-11-09 | 南京信息工程大学 | A kind of cell atypia automatic grading method tactful based on degree of depth study and combination |
CN106202968A (en) * | 2016-07-28 | 2016-12-07 | 北京博源兴康科技有限公司 | The data analysing method of cancer and device |
-
2016
- 2016-12-23 CN CN201611207518.8A patent/CN107526946B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140025689A1 (en) * | 2012-04-24 | 2014-01-23 | International Business Machines Corporation | Determining a similarity between graphs |
CN102722892A (en) * | 2012-06-13 | 2012-10-10 | 西安电子科技大学 | SAR (synthetic aperture radar) image change detection method based on low-rank matrix factorization |
CN103400143A (en) * | 2013-07-12 | 2013-11-20 | 中国科学院自动化研究所 | Data subspace clustering method based on multiple view angles |
CN103793600A (en) * | 2014-01-16 | 2014-05-14 | 西安电子科技大学 | Isolated component analysis and linear discriminant analysis combined cancer forecasting method |
CN105427296A (en) * | 2015-11-11 | 2016-03-23 | 北京航空航天大学 | Ultrasonic image low-rank analysis based thyroid lesion image identification method |
CN106096654A (en) * | 2016-06-13 | 2016-11-09 | 南京信息工程大学 | A kind of cell atypia automatic grading method tactful based on degree of depth study and combination |
CN106202968A (en) * | 2016-07-28 | 2016-12-07 | 北京博源兴康科技有限公司 | The data analysing method of cancer and device |
Non-Patent Citations (2)
Title |
---|
ANINDYA HALDER ET AL.: ""Semi-supervised fuzzy K-NN for cancer classification from microarray gene expression data"", 《2014 FIRST INTERNATIONAL CONFERENCE ON AUTOMATION, CONTROL, ENERGY AND SYSTEMS (ACES)》 * |
刘晋苏: ""微小型无人直升机航拍动态阴影检测研究"", 《中国优秀硕士学位论文全文数据库 工程科技II辑》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108169728A (en) * | 2018-01-12 | 2018-06-15 | 西安电子科技大学 | Range extension target detection method based on Minkowski distances |
CN109378039A (en) * | 2018-08-20 | 2019-02-22 | 中国矿业大学 | Oncogene based on discrete constraint and the norm that binds expresses spectral-data clustering method |
CN109378039B (en) * | 2018-08-20 | 2022-02-25 | 中国矿业大学 | Tumor gene expression profile data clustering method based on discrete constraint and capping norm |
CN109671468A (en) * | 2018-12-13 | 2019-04-23 | 韶关学院 | A kind of feature gene selection and cancer classification method |
CN109671468B (en) * | 2018-12-13 | 2023-08-15 | 韶关学院 | Characteristic gene selection and cancer classification method |
CN109903166A (en) * | 2018-12-25 | 2019-06-18 | 阿里巴巴集团控股有限公司 | A kind of data Risk Forecast Method, device and equipment |
CN109903166B (en) * | 2018-12-25 | 2024-01-30 | 创新先进技术有限公司 | Data risk prediction method, device and equipment |
Also Published As
Publication number | Publication date |
---|---|
CN107526946B (en) | 2021-07-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111860612A (en) | Unsupervised hyperspectral image hidden low-rank projection learning feature extraction method | |
Lee et al. | A new clustering validity index for arbitrary shape of clusters | |
Shen et al. | Multi defect detection and analysis of electron microscopy images with deep learning | |
Drab et al. | Clustering in analytical chemistry | |
CN107526946A (en) | Merge the gene expression data cancer classification method of self study and low-rank representation | |
Gong et al. | Using multi-label classification to improve object detection | |
Tahir et al. | Protein subcellular localization of fluorescence microscopy images: employing new statistical and Texton based image features and SVM based ensemble classification | |
Yang et al. | Active matting | |
CN104820841B (en) | Hyperspectral classification method based on low order mutual information and spectrum context waveband selection | |
CN106203483A (en) | A kind of zero sample image sorting technique of multi-modal mapping method of being correlated with based on semanteme | |
CN108664986B (en) | Based on lpNorm regularized multi-task learning image classification method and system | |
Zhang et al. | Convolutional memory network-based flotation performance monitoring | |
Liu et al. | SRAS‐net: Low‐resolution chromosome image classification based on deep learning | |
Ahn et al. | Searching similar weather maps using convolutional autoencoder and satellite images | |
Yang et al. | Stacking-based and improved convolutional neural network: a new approach in rice leaf disease identification | |
Chen et al. | BiShuffleNeXt: a lightweight bi-path network for remote sensing scene classification | |
Liu et al. | Research on cassava disease classification using the multi-scale fusion model based on EfficientNet and attention mechanism | |
Zhang et al. | Application and evaluation of deep neural networks for airborne hyperspectral remote sensing mineral mapping: a case study of the Baiyanghe uranium deposit in northwestern Xinjiang, China | |
Li et al. | Adaptive mask sampling and manifold to Euclidean subspace learning with distance covariance representation for hyperspectral image classification | |
Wang et al. | Classification and extent determination of rock slope using deep learning | |
CN110287970B (en) | Weak supervision object positioning method based on CAM and covering | |
Gupta et al. | Deep learning for morphological identification of extended radio galaxies using weak labels | |
Dong et al. | ESA-Net: An efficient scale-aware network for small crop pest detection | |
CN114170625A (en) | Context-aware and noise-robust pedestrian searching method | |
Li et al. | HTDFormer: Hyperspectral Target Detection Based on Transformer With Distributed Learning |
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 |