Feature selection based on dragonfly optimization algorithm and its improved for big data classification

Feature selection, is generally achieved by combining an optimization algorithm with a classifier. Dragonfly Algorithm (DA) is a recent swarm intelligence algorithm that mimics the behavior of the dragonflies. Crossover and mutation operators, by changing population, can be efficient in improving the algorithm. In the DA, the initial population is randomly generated and this problem can hinder the achievement of optimal results. Hence, we used chaos theory and created a chaotic population. Results showed that the proposed IDA outperforms traditional DA. For big data classification, the results showed that using term frequency-inverse document frequency (TF-IDF) with proposed algorithm for feature selection is more accurate than using TF-IDF alone.

کلیدواژه ها:

Feature selection ، Dragonfly algorithm ، Optimization ، Big data ، Classification.