Refinement of random forest

• Main Author: Deepika, Mathiyazhagan
• Other Authors: Ponnuthurai N. Suganthan
• Format: Theses and Dissertations
• Language: English
• Published: 2018
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/76335
• Institution: Nanyang Technological University

Random Forest is one of the most popular Machine learning algorithms. It is an ensemble of decision trees and each tree is built using an injection of randomness. The aim of this dissertation: “REFINEMENT OF RANDOM FOREST” is to develop a refined random forest algorithm using Random Vector Functional Link network as a split function to improve the performance. Random Forest has been successfully used in many data mining and computer vision tasks. Despite its immense success, it employs a greedy learning algorithm where locally-optimal decisions are made at each node. The progress of decision making at each node in random forest has been improvised by adapting Random vector functional link network. The random vector functional link network is used to split the decision nodes into two sub-nodes. The Refined Random forest algorithm has better performance as verified in extensive experiments.