Pattern-based Data-Classification Technique
Mostafa A. Salama, A. Hasanen, Ahmed Fahmy · The International Conference on Electrical Engineering/The International Conference on Electrical Engineering · 2010
This paper presents a novel model of a supervised machine learning approach forclassification of a dataset. The model depends on a feature selection (dimensionalityreduction) method that is based on pattern-based subspace clustering technique. Thenthis clustering technique is applied to the dataset to perform the classification of thedata. This approach is a non-statistical technique that supports most of the requirementsthat have been discussed recently like dimensionality reduction using multivariatefeature selection method, threshold independence and handling of missing data. Theapproach tends to handle these requirements altogether which not the case in otherclassification models as discussed in this paper. Another distinguishing point in thismodel is its dependence on the variation of the values of relative features amongdifferent objects. Experimental results on synthetic and real datasets show that approachoutperforms the existing methods in both efficiency and effectiveness.