DocumentCode
2994619
Title
A sequential distribution free pattern classification procedure
Author
Poage, J.L.
Author_Institution
Harvard University, Cambridge, Massachusetts
fYear
1970
fDate
7-9 Dec. 1970
Firstpage
64
Lastpage
64
Abstract
A sequential distribution free pattern classification procedure is presented to classify unknown samples into one of two inseparable classes for the case where the underlying probability density functions of the classes are unknown. The algorithm is formed from training sets of known classification. An estimate (on an expected value basis) is given of the probability of making an error in classification by using order statistics.
Keywords
Electroencephalography; Pattern classification; Physics; Probability; Statistics; Testing; Welding;
fLanguage
English
Publisher
ieee
Conference_Titel
Adaptive Processes (9th) Decision and Control, 1970. 1970 IEEE Symposium on
Conference_Location
Austin, TX, USA
Type
conf
DOI
10.1109/SAP.1970.269957
Filename
4044612
Link To Document