Technical Session - 4

Title:            Design and Development of an Enhanced Random Forest Method to Reduce the Attributes
Authors:       A. Malathi, Gandhimathi
Abstract:     This paper explains the enhancement of Random forest method to reduce the attributes and to make the efficient usage of IDS. The Random Forest is a new ensemble algorithm for classification. This Random Forest uses ensemble unpruned classification or a regression algorithm and which generated many decision trees. Random Forest uses both bagging and boosting as a successful approach for tree building. It is a collection of tree predictors with the same distribution for all trees in the forests. It builds many trees which minimizes the classification errors based on a bootstrap sample of training data from the original dataset using a tree classification algorithm. Random Forest is a most successful method for pre-processing the dataset. The dataset contains many unwanted and irrelevant attributes. This enhanced Random Forest is used to classify the intrusions detection. Pre-process is the process of keeping the dataset ready for the process with necessary attributes for its process.

Keywords:  Random Forest, Pre-processing and Attribute Reduction
Pages:        197-200

Download Full Paper