Enhancing E--Filtering Based on mail Filtering Based on mail Filtering Based on mail Filtering Based on GRF GRF GRF GRF
S. A. El Hafeez, R. El-Awady · 2014
The inferring of insignificant and repetitive features in the d ataset can bring about poor expectations and misclassification process. Subsequently, selecting applicable feature subsets can help decrease the computational cost of feature measurement, accelerate learning process and enhance model interpretability. Feature selection is an issue of worldwide computing optimization in machine learning in which subsets of relevant features are chosen to acknowledge powerful learning models. Rough sets Method in classification has demonstrated wasteful in its failure to deliver accurate and precise classification results about the large e-mail dataset while it likewise expends a ton of computational resources. In this study, we present GRF- Genetics Rough Filter-a hybrid of Genetic Algorithm-Rough set feature selection technique is developed to optimize the Rough set classification parameters, the prediction accuracy and computation time. Spam assassin dataset was used to validate the performance of the proposed system. GRF showed remarkable improvements over Neural Network, Rough Set and SVM methods in terms of classification accuracy.