The Use of Cluster Computing and Random Forest Algoritm for Flight Delay Prediction

Research in High Performance Computing (HPC) is becoming more and more popular recently. Cluster computing is one type of HPC that can be easily deployed. In the meantime, data mining is topic that have been developing where its enormous advantages are unavoidable. When cluster computing and data mining are consolidated, it will yield an exceptionally amazing technology whereby the handling time in data mining can be sped up by utilizing the strength of cluster processing. The expense of creating cluster computing is relatively cheaper contrasted with purchasing a PC with high specs like server PC or even supercomputer. In this paper, we simulate the experiment of cluster computing in virtual environment. The objective of this research is to assess the performance of data mining algorithm (random forest) when executed in cluster environment. In the simulation, we use 1 master and 3 workers node. Those 4 nodes have the same hardware specifications. The result shows that this simulation can speed up random forest performance up to 35.8%. Moreover, the computing performance in cluster environment can be accelerated by attaching more nodes to the cluster with better hardware spesifications