A practical approach to feature selection
ML92 Proceedings of the ninth international workshop on Machine learning
Tolerating noisy, irrelevant and novel attributes in instance-based learning algorithms
International Journal of Man-Machine Studies - Special issue: symbolic problem solving in noisy and novel task environments
Estimating attributes: analysis and extensions of RELIEF
ECML-94 Proceedings of the European conference on machine learning on Machine Learning
Selection of relevant features and examples in machine learning
Artificial Intelligence - Special issue on relevance
Reduction Techniques for Instance-BasedLearning Algorithms
Machine Learning
Rough Sets: Theoretical Aspects of Reasoning about Data
Rough Sets: Theoretical Aspects of Reasoning about Data
Correlation-based Feature Selection for Discrete and Numeric Class Machine Learning
ICML '00 Proceedings of the Seventeenth International Conference on Machine Learning
ECCBR '02 Proceedings of the 6th European Conference on Advances in Case-Based Reasoning
Remembering to forget: a competence-preserving case deletion policy for case-based reasoning systems
IJCAI'95 Proceedings of the 14th international joint conference on Artificial intelligence - Volume 1
Rough set based approaches to feature selection for Case-Based Reasoning classifiers
Pattern Recognition Letters
| Hi-index | 0.00 |
Case-Based Reasoning systems usually retrieve cases using a similarity function based on K-NN or some derivatives. These functions are sensitive to irrelevant or noisy features. Weighting methods are used to extract the most important information present in the knowledge and determine the importance of each feature. However, this knowledge, can also be incorrect, redundant and inconsistent. In order to solve this problem there exist a great number of case reduction techniques in the literature. This paper analyses and justifies the relationship between weighting and case reduction methods, and also analyses their behaviour using different similarity metrics.We have focused this relation on Rough Sets approaches. Several experiments, using different domains from the UCI and our own repository, show that this integration maintain and even improve the performance over a simple CBR system and over case reduction techniques. However, the combined approach produces CBR system decrease if the weighting method declines its performance.