International Journal of Computer Applications |
Foundation of Computer Science (FCS), NY, USA |
Volume 184 - Number 43 |
Year of Publication: 2023 |
Authors: Ojie Deborah Voke, Akazue Maureen, Imianvan Anthony |
10.5120/ijca2023922533 |
Ojie Deborah Voke, Akazue Maureen, Imianvan Anthony . A Framework for Feature Selection using Data Value Metric and Genetic Algorithm. International Journal of Computer Applications. 184, 43 ( Jan 2023), 14-21. DOI=10.5120/ijca2023922533
Most organizations analyse input data to develop an accurate description or model using the features present in the data. There have been huge amount of generated data in the big data ecosystem which demand better and efficient ways to distil high utility or value from it so as to compliment decision makers in making recommendations and decisions. Before applying classification algorithm, relevant features are selected by a suitable feature selection algorithm. Data Value Metric (DVM) is an information theoretic measure based on the notion of mutual information which has been shown to be a good metric to validate the quality and utility of data in a big data ecosystem and in traditional data. Data Value Metric (DVM) suffers from local minima and loss of diversity in the population since it is using forward selection search strategy, however, hybridizing it with Genetic Algorithm is hoped to overcome the problem of local minima as there would be a blend of evolutionary search to ensure a balance between exploration and exploitation of the search space. This paper proposed the hybrid model of Genetic Algorithm and Data Value Metric (DVM) as an information theoretic metric for quantifying the quality and utility for feature selection which can be applied to traditional data.