http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-109961088-A
Outgoing Links
Predicate | Object |
---|---|
assignee | http://rdf.ncbi.nlm.nih.gov/pubchem/patentassignee/MD5_f68ddaa0ddd894016af217422d657c22 |
classificationCPCInventive | http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06F18-22 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06F18-24143 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06F18-214 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06F18-213 |
classificationIPCInventive | http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06K9-62 |
filingDate | 2019-02-13-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
inventor | http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_6494af62480de648b3f600614d566373 http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_7c126fb5ea9562668aa0a62d148618fb http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_f59981eaeccacd9706a320a711ad299a http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_722d0339865e6e73e99187ae06bdf9c5 |
publicationDate | 2019-07-02-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
publicationNumber | CN-109961088-A |
titleOfInvention | Unsupervised nonlinear adaptive manifold learning method |
abstract | The invention discloses an unsupervised nonlinear adaptive manifold learning method. An unsupervised nonlinear adaptive manifold learning method of the present invention includes: expanding neighboring points; merging neighboring points; according to the above, defining an objective function: where α is a trade-off parameter, and the algorithm can be flexibly adjusted by adjusting the value of α A balance between these two considerations in consider local topological relations; x j ∈ MNN(x i ) means that x j is the nearest neighbor of x i after using the adaptive neighbor method above; the reconstruction weight matrix W is obtained by optimizing the following problem. Beneficial effects of the present invention: the algorithm skillfully combines the advantages of LLE and isomap algorithms, and considers local and global features at the same time, and can perform comprehensive and effective feature extraction on high-dimensional data. |
priorityDate | 2019-02-13-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
type | http://data.epo.org/linked-data/def/patent/Publication |
Incoming Links
Total number of triples: 21.