Automatic Generation of Epsilon (Eps) Value for DBSCAN Using Genetic Algorithms

Cluster analysis or Clustering is considered a very important technique in the field of Data Mining. There are many approaches to clustering i.e., partitioning-based, hierarchal based and density based. In such approaches, many algorithms have been developed. One of the most common approaches in clustering algorithms is based on an examination of cluster density. Although density-based algorithms employ a variety of techniques, one of the most often referenced in the academic literature is the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm. The DBSCAN algorithm is used in a wide range of applications. This is due to the fact that this algorithm can find clusters of arbitrary shapes and sizes and its' robustness to noises which contributes a great deal to its success. However, there is a crucial problem with the correct choice of its two input parameters values, namely, the neighborhood radius (Eps) and the MinPts. The purpose of this research is to address the fundamental challenge in employing the clustering algorithm, which is to determine the input parameter (Eps) value automatically, and that by introducing a novel hybrid approach that combines the Genetic Algorithm (GA) with the DBSCAN algorithm, named GA-DBSCAN, to automatic determine the optimal value for the Eps parameter.