54 search results - page 5 / 11 » An Empirical Evaluation of Bagging and Boosting |
AIPRF
2007
13 years 9 months ago
2007
This paper presents experiments on classifying web pages by genre. Firstly, a corpus of 1539 manually labeled web pages was prepared. Secondly, 502 genre features were selected ba...
KDD
2008
ACM
14 years 7 months ago
2008
ACM
We propose a family of novel cost-sensitive boosting methods for multi-class classification by applying the theory of gradient boosting to p-norm based cost functionals. We establ...
CIVR
2004
Springer
14 years 25 days ago
2004
Springer
Our work in content-based image retrieval (CBIR) relies on content-analysis of multiple representations of an image which we term multiple viewpoints or channels. The conceptual id...
SETN
2004
Springer
14 years 24 days ago
2004
Springer
Abstract. Single classifiers, such as Neural Networks, Support Vector Machines, Decision Trees and other, can be used to perform classification of data for relatively simple proble...
FGR
2006
IEEE
14 years 1 months ago
2006
IEEE
We examined the effectiveness of using Haar features and the Adaboost boosting algorithm for FACS action unit (AU) recognition. We evaluated both recognition accuracy and processi...