Feature Selection using Parallel Cuckoo Algorithm with Naïve Bayes Classifier based on Two Different Strategies

摘要

Feature selection is the important step in data preparing phase to reduce complexity of dataset and increase the performance of the classification. In this paper, we proposed Cuckoo Feature Selection (CFS) algorithm, which is based on the natural behavior of cuckoo birds. Moreover, two different classification strategies of naïve Bayes were combined with the proposed algorithm and have been experimented on various datasets from the UCI repository, and the proposed algorithm were compared with previous work to test its performance. In conclusion, this work provides empirical evidence that the proposed CFS algorithm with both classifiers was able to perform satisfactorily over UCI datasets.