DocumentCode
3450904
Title
Learning mixtures of Gaussians
Author
Dasgupta, Sanjoy
Author_Institution
California Univ., Berkeley, CA, USA
fYear
1999
fDate
1999
Firstpage
634
Lastpage
644
Abstract
Mixtures of Gaussians are among the most fundamental and widely used statistical models. Current techniques for learning such mixtures from data are local search heuristics with weak performance guarantees. We present the first provably correct algorithm for learning a mixture of Gaussians. This algorithm is very simple and returns the true centers of the Gaussians to within the precision specified by the user with high probability. It runs in time only linear in the dimension of the data and polynomial in the number of Gaussians
Keywords
learning systems; probability; local search heuristics; mixtures of Gaussians learning; statistical models; Astrophysics; Clustering algorithms; Electrical capacitance tomography; Gaussian processes; Geology; History; Probability; Psychology; Read only memory; Statistics;
fLanguage
English
Publisher
ieee
Conference_Titel
Foundations of Computer Science, 1999. 40th Annual Symposium on
Conference_Location
New York City, NY
ISSN
0272-5428
Print_ISBN
0-7695-0409-4
Type
conf
DOI
10.1109/SFFCS.1999.814639
Filename
814639
Link To Document