Some datasets require highly complicated fuzzy models for the best knowledge ability. The complicated models mean poor intelligibility for humans and more calculations for machines. Thus often the model are artificially simplified to save the interpretability. The simplified model (with less rules) have higher error for knowledge generalisation. In order to improve knowledge generalisation in simplified models it is convenient to reduce the complexity of data by reduction of noise in the datasets. The paper presents the algorithm for noise removal based on modified discrete convolution is crisp domain. The experiments reveal that the algorithm can improve the generalisation ability for simplified model of highly complicated data.
CITATION STYLE
Simiński, K. (2009). Data noise reduction in neuro-fuzzy systems. Advances in Intelligent and Soft Computing, 57, 187–194. https://doi.org/10.1007/978-3-540-93905-4_23
Mendeley helps you to discover research relevant for your work.