Performance Analysis of K Means Clustering Algorithms for mMTC Systems
Haesik Kim
Abstract
In 5G, we generate a huge amount of data everyday due to high capacity network systems. Many research groups paid attention to machine learning algorithms in order to deal with big data and massive connection. The mMTC systems are one of key 5G applications. It requires massive connection. Clustering is one of key research challenges to design mMTC systems. K-means clustering algorithm is one of the simplest unsupervised machine learning algorithms. The purpose of this algorithm is to find a cluster in data by iteratively minimizing the measure between the cluster centre of the group and the given observation. In this paper, K means clustering algorithms are applied for mMTC clustering problem. New metrics for clustering mMTC devices are proposed. Their performances are investigated and analyzed under the given simulation configuration.