http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-107194428-A
Outgoing Links
Predicate | Object |
---|---|
assignee | http://rdf.ncbi.nlm.nih.gov/pubchem/patentassignee/MD5_4cdacda0f306eef1b23ae93bf6a65b13 |
classificationCPCInventive | http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06F18-214 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06F18-24147 |
classificationIPCInventive | http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06K9-62 |
filingDate | 2017-05-26-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
inventor | http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_d1191e2208a652a46c32f64c49b52c5a http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_214fcd6e902dfac5b5d5bffd7ca9a430 |
publicationDate | 2017-09-22-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
publicationNumber | CN-107194428-A |
titleOfInvention | An Integrated Self-Training Method Based on Neighbor Density and Semi-Supervised KNN |
abstract | The invention discloses an integrated self-training method based on the neighbor density and semi-supervised KNN, which uses the neighbor density method to select the initialized marked samples, avoids the k neighbor samples around the marked samples from becoming marked candidate sets, and makes the initialized marked samples The distance between samples should be dispersed as much as possible to better reflect the original spatial structure of samples, and at the same time, the sample with the highest density in the candidate set of labeled samples should be selected as the labeled sample. In order to improve the performance of data clipping, semi-supervised KNN is used instead of WKNN to make up for the fact that when WKNN does data clipping, it only considers the impact of labeled samples on the sample category to be tested, and does not use the unlabeled samples around the sample to be tested. Problem, in UCI The comparative experiments on the dataset verify the effectiveness of the proposed algorithm. |
isCitedBy | http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-110008323-B http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-110008323-A http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-112348108-A http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-113111969-B http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-107818175-B http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-113111969-A http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-107818175-A |
priorityDate | 2017-05-26-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
type | http://data.epo.org/linked-data/def/patent/Publication |
Incoming Links
Predicate | Subject |
---|---|
isDiscussedBy | http://rdf.ncbi.nlm.nih.gov/pubchem/substance/SID415828001 http://rdf.ncbi.nlm.nih.gov/pubchem/compound/CID130640 |
Total number of triples: 22.