In depth Coverage and Analysis of Information Fusion Technique (with Enhanced Algorithm for Feature Selection with Multiple Classifier System) for Data Mining
Amit Ganatra1, Y P Kosta2
1Amit Ganatra, Department of Computer Engineering, Charotar Institute of Technology (Faculty of Technology and Engineering), Charotar University of Technology Changa, Anand, Gujarat, India.
2Y P Kosta, Marwadi Education Foundation’s Group of Institutions, Rajkot, India.
Manuscript received on November 21, 2011. | Revised Manuscript received on November 30, 2011. | Manuscript published on January 05, 2012. | PP: 66-73 | Volume-1 Issue-6, January 2012. | Retrieval Number: F026911151/2012©BEIESP
Open Access | Ethics and Policies | Cite
© The Authors. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: The main objective of Information Fusion techniques in Data Mining is to show that final information generated to be of superior quality and more meaningful, than the information available from the actual (primary) sources. Fusion, by definition, requires a qualitative difference between the final output and the output of the original sources. Information fusion is the process of acquisition, filtering, correlation and integration of relevant information from various sources into one representational format that is appropriate for deriving decisions regarding the interpretation of the information. In theory, the fusion of redundant information from different sources can reduce redundancy and overall uncertainty and thus increase the accuracy of the system. The fusion can be performed on three levels: raw data level, feature level, or decision level. This paper presents a novel idea of a multiple (ensemble) classification (classifier) system with feature selection where Neural Networks (Multilayer Feed-forward Networks with Back Propagation learning) are boosted for scalable (High Dimensional) datasets. The method uses Genetic Algorithms for Feature Selection with various Evaluation Techniques (Evaluators) like subset evaluation, consistency subset evaluation and wrapper subset approaches to enhance the performance of the feature selection and overall system.
Keywords: Classification, Multiple Classifier Pre-processing, Training, Testing, Feature Selection, Attribute Selected Classifier (ASC).