CN110993121A - Drug association prediction method based on double-cooperation linear manifold - Google Patents
Drug association prediction method based on double-cooperation linear manifold Download PDFInfo
- Publication number
- CN110993121A CN110993121A CN201911239229.XA CN201911239229A CN110993121A CN 110993121 A CN110993121 A CN 110993121A CN 201911239229 A CN201911239229 A CN 201911239229A CN 110993121 A CN110993121 A CN 110993121A
- Authority
- CN
- China
- Prior art keywords
- drug
- incidence
- manifold
- matrix
- target
- 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
- 239000003814 drug Substances 0.000 title claims abstract description 67
- 238000000034 method Methods 0.000 title claims abstract description 65
- 229940079593 drug Drugs 0.000 title claims abstract description 60
- 239000011159 matrix material Substances 0.000 claims abstract description 60
- 108090000623 proteins and genes Proteins 0.000 claims abstract description 36
- 102000004169 proteins and genes Human genes 0.000 claims abstract description 35
- 102000038037 druggable proteins Human genes 0.000 claims abstract description 16
- 108091007999 druggable proteins Proteins 0.000 claims abstract description 16
- 238000007418 data mining Methods 0.000 claims abstract description 4
- 230000009977 dual effect Effects 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 3
- 230000000452 restraining effect Effects 0.000 claims description 3
- 238000007877 drug screening Methods 0.000 claims description 2
- 230000006870 function Effects 0.000 description 7
- 230000008901 benefit Effects 0.000 description 4
- 238000000354 decomposition reaction Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 230000008406 drug-drug interaction Effects 0.000 description 3
- 238000005065 mining Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000004850 protein–protein interaction Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H70/00—ICT specially adapted for the handling or processing of medical references
- G16H70/40—ICT specially adapted for the handling or processing of medical references relating to drugs, e.g. their side effects or intended usage
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
Abstract
A medicine correlation prediction method based on double-cooperation linear manifold belongs to the field of data mining and biological information, and comprises the following steps: 1) constructing an initial target incidence matrix between drug-drug nodes according to incidence relation data between drugs; 2) constructing a drug-target protein and a target protein-target protein node auxiliary incidence matrix according to incidence relation data among drug target proteins and incidence relation data among target proteins; 3) acquiring the manifold of the initial target incidence matrix and the auxiliary matrix as input, and constructing a dual-cooperation linear manifold learning model; 4) according to the consistency principle, target matrix information is enriched through iterative updating, higher-grade association is obtained, association relation between the two medicines is considered to exist, and the medicine association prediction task is completed. The invention adopts the manifold to measure the correlation of the data and adopts the cooperative learning to fully utilize the consistency in the network. The method can effectively improve the accuracy of prediction and is suitable for drug-drug correlation prediction.
Description
Technical Field
The invention relates to the field of data mining and biological information, in particular to a method for performing association prediction on nodes in a heterogeneous network.
Background
In recent years, the problem of link prediction in heterogeneous networks has received a great deal of attention. Many problems associated with this are being studied extensively, including friend recommendations in social networks, protein-protein interaction predictions, and airline network restructuring, among others. Currently, many methods have been proposed to improve the link prediction method by using the topology and interconnection of heterogeneous networks, which are mainly classified into three categories: similarity-based algorithms, path-based algorithms and matrix decomposition-based algorithms. In the similarity-based method, the correlation between nodes is measured by calculating a similarity score. These algorithms, although less computationally expensive, do not have high prediction accuracy because they do not take full advantage of the global structure in known networks. Path-based algorithms typically make use of network topology and node properties for link prediction. However, these methods are computationally expensive to learn the structure globally from the network. Matrix factorization based algorithms can extract potential features from known networks for link prediction. However, existing decomposition-based algorithms do not fully utilize the information in the auxiliary network nor retain valid information during feature integration.
More recently, manifold learning for link prediction has become increasingly popular in the field of machine learning and pattern recognition. The basic idea of manifold learning is to project data from an original high-dimensional space to another low-dimensional space, so that more potential information can be learned, reproducing the basic structural features in the original data. It is superior to algorithms that consider only the original feature space. After size transformation, redundant features in the original feature space are deleted, and the interrelation between nodes is reconstructed to better characterize the similarity of the nodes. Manifold learning is widely used in knowledge representation models due to its effectiveness for data distance measurements. Wan et al, in Feature extraction using two-dimensional maximum embedding difference, propose a two-dimensional maximum embedding difference (2d med) method that combines graph embedding and difference criterion techniques for image Feature extraction. Zhang et al, man's modified customized matrix regularization for drug-drug interaction prediction, introduced a drug-feature based method of regularization of Manifold regularization that projects drugs in an interaction space into a low dimensional space to predict drug-drug interactions.
In summary, none of the existing heterogeneous network link prediction methods considers the consistency between the target network and the auxiliary network, resulting in lower prediction performance.
Disclosure of Invention
The invention mainly aims to solve the problem that the existing medicine correlation prediction technology cannot fully utilize the correlation relation structure information between heterogeneous nodes, and provides a method for performing medicine correlation prediction on nodes in a heterogeneous network based on the manifold consistency between a target network and an auxiliary network. The invention realizes a drug correlation prediction method based on double cooperative linear manifold, optimizes the node similarity by cooperatively using the manifold consistency embedded between a target network and an auxiliary network, and achieves good correlation prediction effect.
Technical scheme of the invention
A drug association prediction method based on double-cooperation linear manifold is disclosed, the prediction result of the method can be applied to data mining or drug screening, and the method specifically comprises the following steps:
step 1: constructing an initial target incidence matrix between drug-drug nodes according to incidence relation data between drugs;
step 2: constructing a drug-target protein and a target protein-target protein node auxiliary incidence matrix according to incidence relation data among drug target proteins and incidence relation data among target proteins;
and step 3: acquiring the manifold of the initial target incidence matrix and the auxiliary incidence matrix as input, and constructing a dual-cooperation linear manifold learning model;
and 4, step 4: according to the consistency principle, target incidence matrix information is enriched through iterative updating, higher-grade incidence is obtained, incidence relation between the two medicines is considered to exist, and a medicine incidence prediction task is completed.
Further, the method of the present invention relates to three kinds of correlation matrices, which are respectively defined as follows: and respectively constructing incidence matrixes of the medicines, the target proteins and the target proteins according to incidence relation data among the medicines, the target proteins and the target proteins, marking 1 if known incidence information exists in any incidence, and marking 0 if the known incidence information exists in any incidence.
The dual-cooperation linear manifold learning model consists of a consistency manifold information item, a priori knowledge constraint item and a sparse constraint.
Further, the consistent manifold information item is:
the method is used for constraining the manifold enriched according to the two-way knowledge, and cooperatively keeping the difference between the two manifolds to be minimum under the condition of enriching more information as much as possible.
Further, the constraint term of the prior knowledge is as follows:
the method is used for restraining the drug-drug association matrix and the target protein-target protein association matrix obtained in the enrichment process, and keeping the unification with the prior knowledge under the condition of being enriched as much as possible.
Further, the sparse constraint is: for controlling the complexity level.
The invention has the advantages and beneficial effects that:
the invention can better capture the correlation among data points and fully utilize the consistency inside the network to carry out the correlation prediction on the nodes in the heterogeneous network. In addition, the technique employs low rank constraints, and incorporates a priori knowledge to overcome sparsity issues of heterogeneous networks, which can maintain stable performance over networks containing a large number of interactions that are missing or not observed. Compared with the current latest technology, the invention obviously improves the accuracy of predicting unknown links in different practical applications, and the applications comprise a social evaluation network, a gene phenotype association prediction and a drug-drug interaction network, and have certain expansibility.
Drawings
FIG. 1 is a flow chart of a drug association prediction method based on dual collaborative linear manifold.
Fig. 2 is a schematic diagram of a method of expanding local manifolds to global manifolds.
FIG. 3 is a schematic diagram of a dual collaborative linear prevalence model.
FIG. 4 is a schematic representation of a ROC curve.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention are described in further detail below.
Example 1:
a drug association prediction method based on dual cooperative linear manifold is disclosed, the flow chart is shown in figure 1, the model chart is shown in figure 3, the method comprises the following steps:
step 101: constructing a drug-drug node correlation matrix according to the correlation relation data among the drugs captured by the drug bank database;
step 102: introducing auxiliary network data captured by a drug Bank database and an HPRD database, and constructing a drug-target protein internode auxiliary incidence matrix and a target protein-target protein internode auxiliary incidence matrix according to the target protein interrelateddata and the drug target protein interrelateddata;
wherein, in step 101, the drug-drug association matrix is the target network T (758 × 758); in step 102, the drug-target protein internodal correlation matrix is the correlation matrix R (758 × 473), the target protein-target protein correlation matrix is the auxiliary matrix A (473 × 473), and the three networks together form an integrated heterogeneous network, which includes 758 drugs and 473 target proteins. If the network nodes have correlation, the corresponding matrix element is 1, otherwise, the corresponding matrix element is 0;
step 103: acquiring the manifold of the initial target incidence matrix and the auxiliary incidence matrix as input, and constructing a dual-cooperation linear manifold learning model;
the dual-cooperation linear manifold learning model consists of a consistency manifold information item, a priori knowledge constraint item and a sparse constraint.
Wherein, the consistent manifold information items in the collaborative linear manifold model are as follows: the method is used for constraining the manifold enriched according to the two-way knowledge, and cooperatively keeping the difference between the two manifolds to be minimum under the condition of enriching more information as much as possible.
The prior knowledge constraint term is: the method is used for restraining the drug-drug association matrix and the target protein-target protein association matrix obtained in the enrichment process, and keeping the unification with the prior knowledge under the condition of being enriched as much as possible.
The sparse constraint is as follows: for controlling the complexity level.
Wherein, the linear manifold learning refers to a local linear embedding method and a sparse subspace clustering method.
Wherein the locally linear embedding assumes a data point X sampled from the data set XiAnd phi (I) denotes its k neighbors. Each point xiCan be approximated as a linear combination of the midpoints of phi (I). Coefficient unit omegaijIs used for reconstructing XiPoint X ofjE.g., phi (i), as shown in equation (1)
Further, sparse subspace clustering is integrated with manifold, extending the local neighborhood to the global space to fully utilize the advantages of manifold learning, as shown in fig. 2. Known drug-drug associations are embedded as a priori knowledge to reconstruct drug target protein interconnection matrices, and fixed weight matrices W are integrated to control the effects of the a priori knowledge. The prior constraint to get T is expressed as follows:
wherein T is(0)Is a adjacency matrix of known drug association information, and is marked as 1 for the known association information, and is 0 otherwise. In order to retain the known drug association relationship, when there is association between drugs, the value of W is 0.8, and when there is no association, the value of W is 0.2.
Using a priori knowledge as a constraint of the loss function (a priori constraint) such thatThe manifold constraint is more flexible, thus obtaining a robust solution. Due to the sparsity of the target network, the kernel norm is adopted to carry out low-rank constraint T, which is expressed as T(*). The loss function for linear manifold learning is thus derived as follows:
where α and γ are hyper-parameters for coordinating the weights of the a priori constraint and the low rank, low order constraint.
Similarly, information about the deletion correlation between target proteins can also be inferred by linear manifold learning. A similar loss function on the auxiliary network a is derived as follows:
the idea of local invariance states that the results obtained from learning a linear manifold from different directions have similarities. Replacing the manifold constraint of the drug associated network and the manifold constraint of the target protein associated network with the cooperative constraint of the drug and the target protein network to obtain a final loss function of
Wherein W1And W2The prior knowledge weighting matrixes of the drug network and the target protein network respectively, α, β and gamma are two hyper-parameters, values of which are 1000, 1000 and 0.5 respectively, and a final drug-drug association matrix is obtained after convergence through collaborative manifold learning, so that potential drug association information is obtained.
Further, the optimization of the collaborative linear manifold learning model specifically comprises:
first, the drug association matrix is updated by fixing the target protein association matrix, and the loss function of the drug association matrix T is expressed as:
According to the gradient optimization method, the iterative equation of the obtained drug correlation matrix is
Ignoring the irrelevant term f of T (T)(k-1)) Will be
Is shown as
The iterative equation of T obtained by the SVT method is
First, the update step size of T is calculated. In combination with the conditions of the optimization, the method,
the step updating mode is as follows: when in use
When the temperature of the water is higher than the set temperature,
wherein the content of the first and second substances,
after the update is completed, the final result is obtained
I.e. the step size of the kth iteration of the target network T.
Also, fixed drug association using gradient optimization methodsThe matrix updates the target protein incidence matrix to obtain the step length of the Kth iteration of the auxiliary network A
Further, combining
The update equations for T and A are as follows:
step 104: according to the consistency principle, target matrix information is enriched through iterative updating, and if the relation value between two certain medicines is larger than a specific threshold value, the medicines are considered to be related, so that the final prediction result is obtained.
Iteratively updating T and A until the model converges, and solving a final similarity matrix
To predict missing link information in the target network T.
Wherein, PikThe element value corresponding to the ith row and the kth column in the matrix PP represents the correlation between the drug i and the target protein k, and the larger the value, the stronger the correlation. When P is equal to 0.8 by setting the threshold MikWhen the protein is more than M, the drug i is considered to be related to the target protein k.
In conclusion, through the processing of the steps 101 to 104, the method realizes modeling and solving of the drug-target protein association relation, and achieves a good drug-target protein association prediction effect.
The feasibility of the method of the present invention was verified in the following experiments, and the examples of the present invention were verified using the true drug-drug, drug-target protein and target protein-target protein associations captured in the drug bank database and the HPRD database.The two data sets are derived fromhttps://www.drugbank.caAndhttp://hprd.org/index_htmlthe public data set of (1). The data sets and experimental results will be described separately below.
The detailed data content comprises three related data of FDA-certified drug-drug action (DDI), drug-target protein action (DTI) and target protein-target protein association (PPI). Wherein the DDI dataset contains a total of 828 drugs and 14746 drug-drug associations, the DTI dataset contains 473 target proteins and 2416 drug-target proteins, and the PPI dataset contains 1098 target protein-target protein associations. After preprocessing and aligning the three types of correlation data, we finally obtain a 758 × 758 known DDI correlation matrix, a 758 × 473 DTI correlation matrix and a 473 × 473 PPI correlation matrix, wherein there are 5926 pairs of correlations in the DDI correlation matrix, 2416 pairs of correlations in the DTI correlation matrix, and 616 pairs of correlations in the PPI correlation matrix. Corresponding to T, R, a in the bi-cooperative linear manifold model.
In order to evaluate the performance of a dual Collaborative linear manifold model (CLML) proposed by the method of the present invention, neighbor set information is adopted as NSI [1], a matrix completion Link Prediction method is adopted as MCLP [2], and a matrix decomposition method MRMF [3] based on manifold regularization is adopted, wherein the coefficient lambda of a balance constraint term is set to 0.1 for the matrix completion Link Prediction method, three parameters α and β and gamma are respectively set to 0.1,0.1 and 0.3 for the matrix completion Link Prediction method, and the coefficient gamma of the constraint term is set to 0.25 and mu is set to 1 for the matrix decomposition method based on manifold regularization.
The method of the invention selects a mode of using 5-fold cross test to divide a training set and a testing set, and uses AUC [4] as an evaluation index, wherein the AUC is defined as follows:
area under the ROC curve, as shown in fig. 4. The ROC curve is a test subject working characteristic curve which is drawn by taking the true positive rate as the ordinate and the false positive rate as the abscissa according to a series of different two classification modes. The larger the graphic area, the better the classification effect.
The experimental results of each method are shown in table 1:
TABLE 1
The experimental results of the CLML algorithm and the benchmark algorithm provided by the method of the invention show that the method based on the dual-cooperation linear manifold model is superior to the traditional NSI method, and the method can better utilize the information in the matrix and improve the model performance by a mode of fully mining potential characteristics through the matrix. The CLML method is superior to all other reference algorithms, shows that the acquisition of data structure information by manifold learning and the further use of auxiliary information by dual-cooperation manifolds all improve the performance of the model, and simultaneously shows that the bidirectional cooperation linear manifold constraint designed by the method is effective on the problem of medicine correlation prediction function module mining.
In conclusion, the drug association prediction method based on the double-cooperation linear manifold, which is provided by the method, is superior to the benchmark algorithm. Further, the linear manifold information between the drug-drug association relationship, the target protein-target protein association relationship and the drug-target protein association relationship is considered, so that the performance of the drug association prediction function module mining task can be improved.
Those skilled in the art will appreciate that the drawings are merely schematic representations of one preferred embodiment.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Reference to the literature
[1]B.Zhu,Y.Xia,An information-theoretic model for link prediction incomplex networks,Sci.Rep.5(2015)13707.
[2]M.Gao,L.Chen,B.Li,et al.,A link prediction algorithm based on low-rank matrix completion,Appl.Intell.48(12)(2018)4531–4550.
[3]W.Zhang,Y.Chen,D.Li,et al.,Manifold regularized matrixfactorization for drug-drug interaction prediction,J.Biomed.Inf.88(2018)90–97.
[4]J.A.Hanley,B.J.McNeil,The meaning and use of the area under areceiveroperating characteristic(ROC)curve,Radiology 143(1)(1982)29–36.
Claims (6)
1. A drug association prediction method based on double-cooperation linear manifold is characterized in that the prediction result of the method can be applied to data mining or drug screening, and the method specifically comprises the following steps:
step 1: constructing an initial target incidence matrix between drug-drug nodes according to incidence relation data between drugs;
step 2: constructing a drug-target protein and a target protein-target protein node auxiliary incidence matrix according to incidence relation data among drug target proteins and incidence relation data among target proteins;
and step 3: acquiring the manifold of the initial target incidence matrix and the auxiliary incidence matrix as input, and constructing a dual-cooperation linear manifold learning model;
and 4, step 4: according to the consistency principle, target incidence matrix information is enriched through iterative updating, higher-grade incidence is obtained, incidence relation between the two medicines is considered to exist, and a medicine incidence prediction task is completed.
2. The dual-coordination linear manifold-based drug association prediction method according to claim 1, wherein the method involves three association matrices, which are respectively defined as follows: and respectively constructing incidence matrixes of the medicines, the target proteins and the target proteins according to incidence relation data among the medicines, the target proteins and the target proteins, marking 1 if known incidence information exists in any incidence, and marking 0 if the known incidence information exists in any incidence.
3. The dual-cooperation linear manifold-based drug correlation prediction method according to claim 1, wherein the dual-cooperation linear manifold learning model is composed of a consistent manifold information item, a priori knowledge constraint item and a sparse constraint.
4. The dual coordinated linear manifold-based drug association prediction method as claimed in claim 3, wherein said consistent manifold information item is:
the method is used for constraining the manifold enriched according to the two-way knowledge, and cooperatively keeping the difference between the two manifolds to be minimum under the condition of enriching more information as much as possible.
5. The method according to claim 3, wherein the prior knowledge constraint term is:
the method is used for restraining the drug-drug association matrix and the target protein-target protein association matrix obtained in the enrichment process, and keeping the unification with the prior knowledge under the condition of being enriched as much as possible.
6. The dual-coordination linear manifold-based drug association prediction method according to claim 3, wherein the sparse constraint is: for controlling the complexity level.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911239229.XA CN110993121A (en) | 2019-12-06 | 2019-12-06 | Drug association prediction method based on double-cooperation linear manifold |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911239229.XA CN110993121A (en) | 2019-12-06 | 2019-12-06 | Drug association prediction method based on double-cooperation linear manifold |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110993121A true CN110993121A (en) | 2020-04-10 |
Family
ID=70090652
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911239229.XA Pending CN110993121A (en) | 2019-12-06 | 2019-12-06 | Drug association prediction method based on double-cooperation linear manifold |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110993121A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112951320A (en) * | 2021-03-03 | 2021-06-11 | 深圳大学 | Biomedical network association prediction method based on ensemble learning |
| CN116153391A (en) * | 2023-04-19 | 2023-05-23 | 中国人民解放军总医院 | Antiviral drug screening method, system and storage medium based on joint projection |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105512511A (en) * | 2015-11-24 | 2016-04-20 | 南开大学 | Multi-non-negative matrix resolution clustering method based on consistency |
| CN106709272A (en) * | 2016-12-26 | 2017-05-24 | 西安石油大学 | Method and system for predicting drug-target protein interaction relationship based on decision template |
| CN107506591A (en) * | 2017-08-28 | 2017-12-22 | 中南大学 | A kind of medicine method for relocating based on multivariate information fusion and random walk model |
-
2019
- 2019-12-06 CN CN201911239229.XA patent/CN110993121A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105512511A (en) * | 2015-11-24 | 2016-04-20 | 南开大学 | Multi-non-negative matrix resolution clustering method based on consistency |
| CN106709272A (en) * | 2016-12-26 | 2017-05-24 | 西安石油大学 | Method and system for predicting drug-target protein interaction relationship based on decision template |
| CN107506591A (en) * | 2017-08-28 | 2017-12-22 | 中南大学 | A kind of medicine method for relocating based on multivariate information fusion and random walk model |
Non-Patent Citations (1)
| Title |
|---|
| WEN ZHANG等 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112951320A (en) * | 2021-03-03 | 2021-06-11 | 深圳大学 | Biomedical network association prediction method based on ensemble learning |
| CN116153391A (en) * | 2023-04-19 | 2023-05-23 | 中国人民解放军总医院 | Antiviral drug screening method, system and storage medium based on joint projection |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109492822B (en) | Air pollutant concentration time-space domain correlation prediction method | |
| CN110347932B (en) | Cross-network user alignment method based on deep learning | |
| Maraziotis | A semi-supervised fuzzy clustering algorithm applied to gene expression data | |
| CN108182260B (en) | Multivariate time sequence classification method based on semantic selection | |
| CN105160400A (en) | L21 norm based method for improving convolutional neural network generalization capability | |
| CN111898703B (en) | Multi-label video classification method, model training method, device and medium | |
| CN113052333A (en) | Method and system for data analysis based on federal learning | |
| CN109741364B (en) | Target tracking method and device | |
| CN112070277A (en) | Hypergraph neural network-based drug-target interaction prediction method | |
| CN113190688A (en) | Complex network link prediction method and system based on logical reasoning and graph convolution | |
| CN114639483A (en) | Electronic medical record retrieval method and device based on graph neural network | |
| CN111916144A (en) | Protein classification method based on self-attention neural network and coarsening algorithm | |
| CN110993121A (en) | Drug association prediction method based on double-cooperation linear manifold | |
| CN114999635A (en) | circRNA-disease association relation prediction method based on graph convolution neural network and node2vec | |
| CN113987236B (en) | Unsupervised training method and unsupervised training device for visual retrieval model based on graph convolution network | |
| CN114358250A (en) | Data processing method, data processing apparatus, computer device, medium, and program product | |
| Zhang et al. | Tensorized topological graph learning for generalized incomplete multi-view clustering | |
| CN113066537B (en) | Compound classification method based on graph neural network | |
| CN115481256A (en) | Inverse relation rotation embedding knowledge representation method and system based on convolution quaternion | |
| CN108304546B (en) | Medical image retrieval method based on content similarity and Softmax classifier | |
| CN110289098B (en) | Risk prediction method based on clinical examination and medication intervention data | |
| CN111612572A (en) | Adaptive local low-rank matrix approximate modeling method based on recommendation system | |
| CN113642592A (en) | Training method of training model, scene recognition method and computer equipment | |
| CN114255865A (en) | Diagnosis and treatment project prediction method based on recurrent neural network | |
| CN117593594B (en) | Brain MRI image classification method, equipment and medium based on consistency alignment |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200410 |