Feature Selection Using Genetic Algorithm for Big Data

Feature selection is a powerful technique for dimensionality reduction and an important step in successful machine learning applications. In the last few decades, data has become progressively larger in both numbers of instances and features which make it harder to deal with the feature selection problem. To cope with this new epoch of big data, new techniques need to be developed for addressing this problem effectively. Nonetheless, the suitability of current feature selection algorithms is extremely downgraded and are inapplicable, when data size exceeds hundreds of gigabytes. In this paper, we introduce a scalable implementation of a parallel feature selection approach using the genetic algorithm that has been done in parallel using MapReduce model. The experimental results showed that the proposed method can be suitable to improve the performance of feature selection.