Title :
Lazy MetaCost Naive Bayes
Author :
Kotsiantis, Sotiris ; Kanellopoulos, Dimitris
Author_Institution :
Univ. of Patras, Patras
Abstract :
This paper firstly provides a review on the various methodologies that have tried to handle the problem of learning from data sets with an unbalanced class distribution. Finally, it presents an experimental study of these methodologies with the local application of Metacost algorithm and it concludes that such a framework can be a more effective solution to the problem.
Keywords :
belief networks; data analysis; learning (artificial intelligence); data sets; lazy MetaCost naive Bayes; machine-learning methods; unbalanced class distribution; Costs; Credit cards; Information technology; Laboratories; Machine learning; Machine learning algorithms; Mathematics; Medical diagnosis; Programming; Training data;
Conference_Titel :
Convergence Information Technology, 2007. International Conference on
Conference_Location :
Gyeongju
Print_ISBN :
0-7695-3038-9
DOI :
10.1109/ICCIT.2007.82
