Bipartite Weighted Graph Access for Optimal Label Prediction
Jyothi Puligadda
Dr. Jyothi Puligadda, Assistant Professor, Department of Mathematics, Anurag Engineering College, An UGC Autonomous Institution Kodad, Suryapet (Telangana), India.
Manuscript received on 25 August 2019 | Revised Manuscript received on 11 September 2019 | Manuscript Published on 17 September 2019 | PP: 1848-1852 | Volume-8 Issue-2S8 August 2019 | Retrieval Number: B11670882S819/2019©BEIESP | DOI: 10.35940/ijrte.B1167.0882S819
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. 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: This article portrayed a novel bipartite weighted graph strategy for Feature optimization for machine learning models. Unlike many of the existing optimization techniques of diverse categories such as evolutionary computation techniques, diversity assessment strategies, the proposal is deterministic approach with minimal computational overhead, which has referred further as Bipartite Weighted Graph Approach for Optimal label prediction (BWG-OLP). The proposed model is about to derive a given feature is optimal or not by the respective feature’s correlation with the records and the correlation with the fellow features. The experimental study has carried on benchmark datasets to estimate the significance of the proposed method.
Keywords: Mutual Information, Bipartite Weighted Edge Graph, LASSO, Particle-Swarm Optimization, Degree of Positive Label Scope.
Scope of the Article: Cryptography and Applied Mathematics