35 search results - page 2 / 7 » A Prototype-driven Framework for Change Detection in Data St... |
NIPS
2001
13 years 6 months ago
2001
Probabilistic mixture models are used for a broad range of data analysis tasks such as clustering, classification, predictive modeling, etc. Due to their inherent probabilistic na...
CIKM
2009
Springer
13 years 8 months ago
2009
Springer
Dynamic data streams are those whose underlying distribution changes over time. They occur in a number of application domains, and mining them is important for these applications....
DEXA
2006
Springer
13 years 8 months ago
2006
Springer
We introduce a classification framework for continuous multivariate stream data. The proposed approach works in two steps. In the preprocessing step, it takes as input a sliding wi...
ICMCS
2007
IEEE
13 years 11 months ago
2007
IEEE
We present a novel genre-independent SVM framework for detecting scene changes in broadcast video. Our framework works on content from a diverse range of genres by allowing sets o...
ICDM
2010
IEEE
13 years 2 months ago
2010
IEEE
Abstract--The problem of data stream classification is challenging because of many practical aspects associated with efficient processing and temporal behavior of the stream. Two s...