Iterative Consistency-Based Feature Selection and Its Application to Nucleotide Sequences of Influenza A Viruses

  • Kouichi Hirata Kyushu Institute of Technology


In this paper, first we formulate a consistency-based feature selection problem as combinatorial optimization problems.
Next, for the purpose of increasing the number of instances explained by the features rather than decreasing the number of features themselves in consistency-based feature selection,
we introduce an iterative consistency-based feature selection and design the algorithm to compute it.
Finally, we apply the method to several nucleotide sequences of influenza A viruses and evaluate the advantage of the method.
Technical Papers (Information and Communication Technology)