dc.contributor.author |
Dastile, X
|
|
dc.contributor.author |
Jäger, G
|
|
dc.contributor.author |
Debba, Pravesh
|
|
dc.contributor.author |
Cho, Moses A
|
|
dc.date.accessioned |
2012-12-05T11:31:46Z |
|
dc.date.available |
2012-12-05T11:31:46Z |
|
dc.date.issued |
2012-11 |
|
dc.identifier.citation |
Dastile, X., Jäger, G., Debba, P. and Cho, M.A. 2012. Combining binary classifiers to improve tree species discrimination at leaf level. Conference Proceedings of the 54th Annual Conference of the South African Statistical Association, Port Elizabeth, 5-9 November 2012 |
en_US |
dc.identifier.uri |
http://www.sasa2012.co.za/
|
|
dc.identifier.uri |
http://hdl.handle.net/10204/6399
|
|
dc.description |
Conference Proceedings of the 54th Annual Conference of the South African Statistical Association, Port Elizabeth, 5-9 November 2012 |
en_US |
dc.description.abstract |
This paper focuses on the discrimination of seven different savannah tree species at leaf level using hyperspectral data. The data is small in size, high-dimensional and shows large within-species variability combined with small between species variability which makes discrimination between the tree species (hereafter referred to as classes) challenging. We focus on two classification methods: K-nearest neighbour and feed-forward neural networks for the discrimination of the classes. For both methods, direct 7-class prediction results in high misclassification rates. We therefore construct binary classifiers for all possible binary classification problems and combine them using Error Correcting Output Codes (ECOC) to form a 7-class predictor. ECOC with 1-nearest neighbour binary classifiers result in no improvement compared to a 1-nearest neighbour 7-class predictor whereas ECOC with neural networks binary classifiers improve accuracy by 10% compared to neural networks 7-class predictor, and error rates become acceptable. |
en_US |
dc.language.iso |
en |
en_US |
dc.relation.ispartofseries |
Workflow;9886 |
|
dc.subject |
Error Correcting Output Codes |
en_US |
dc.subject |
ECOC |
en_US |
dc.subject |
Neural Networks |
en_US |
dc.subject |
K-nearest Neighbour |
en_US |
dc.subject |
Hyperspectral Data |
en_US |
dc.subject |
Binary Classifiers |
en_US |
dc.title |
Combining binary classifiers to improve tree species discrimination at leaf level |
en_US |
dc.type |
Conference Presentation |
en_US |
dc.identifier.apacitation |
Dastile, X., Jäger, G., Debba, P., & Cho, M. A. (2012). Combining binary classifiers to improve tree species discrimination at leaf level. http://hdl.handle.net/10204/6399 |
en_ZA |
dc.identifier.chicagocitation |
Dastile, X, G Jäger, Pravesh Debba, and Moses A Cho. "Combining binary classifiers to improve tree species discrimination at leaf level." (2012): http://hdl.handle.net/10204/6399 |
en_ZA |
dc.identifier.vancouvercitation |
Dastile X, Jäger G, Debba P, Cho MA, Combining binary classifiers to improve tree species discrimination at leaf level; 2012. http://hdl.handle.net/10204/6399 . |
en_ZA |
dc.identifier.ris |
TY - Conference Presentation
AU - Dastile, X
AU - Jäger, G
AU - Debba, Pravesh
AU - Cho, Moses A
AB - This paper focuses on the discrimination of seven different savannah tree species at leaf level using hyperspectral data. The data is small in size, high-dimensional and shows large within-species variability combined with small between species variability which makes discrimination between the tree species (hereafter referred to as classes) challenging. We focus on two classification methods: K-nearest neighbour and feed-forward neural networks for the discrimination of the classes. For both methods, direct 7-class prediction results in high misclassification rates. We therefore construct binary classifiers for all possible binary classification problems and combine them using Error Correcting Output Codes (ECOC) to form a 7-class predictor. ECOC with 1-nearest neighbour binary classifiers result in no improvement compared to a 1-nearest neighbour 7-class predictor whereas ECOC with neural networks binary classifiers improve accuracy by 10% compared to neural networks 7-class predictor, and error rates become acceptable.
DA - 2012-11
DB - ResearchSpace
DP - CSIR
KW - Error Correcting Output Codes
KW - ECOC
KW - Neural Networks
KW - K-nearest Neighbour
KW - Hyperspectral Data
KW - Binary Classifiers
LK - https://researchspace.csir.co.za
PY - 2012
T1 - Combining binary classifiers to improve tree species discrimination at leaf level
TI - Combining binary classifiers to improve tree species discrimination at leaf level
UR - http://hdl.handle.net/10204/6399
ER -
|
en_ZA |