NOISE DETECTION AND ELIMINATION IN DATA PREPROCESSING: EXPERIMENTS IN MEDICAL DOMAINS

Authors: Gamberger a.; Lavrac N.; Dzeroski S.

Source: Applied Artificial Intelligence, Volume 14, Number 2, 1 February 2000 , pp. 205-223(19)

Buy & download fulltext article:

Price: $56.94 plus tax (Refund Policy)

Abstract:

Compression measures used in inductive learners, such as measures based on the minimum description length principle, can be used as a basis for grading candidate hypotheses. Compression-based induction is suited also for handling noisy data. This paper shows that a simple compression measure can be used to detect noisy training examples, where noise is due to random classification errors. A technique is proposed in which noisy examples are detected and eliminated from the training set, and a hypothesis is then built from the set of remaining examples. This noise elimination method was applied to preprocess data for four machine-learning algorithms, and evaluated on selected medical domains.

Language: English

Document Type: Research article

Publication date: 2000-02-01

More about this publication?