Mining Big Data with Decision Tree Algorithm

Decision tree algorithms are very useful approaches in data mining. Indeed, the C4.5 algorithm is a popular data classifier for machine learning. Nowadays there is a wide range of Big Data frameworks such as Hadoop and Apache spark. In this article, we present a comparative study of different implementations of the C4.5 algorithm in a Big Data environment. Then a new approach is proposed for the implementation of the parallel version of the C4.5 algorithm (PC4.5) based on the Apache Spark framework. Experimental results show the superiority of the proposed PC4.5 algorithm over its sequential version and the version on the Hadoop platform.