Optimizing DBSCAN Density-Based Clustering Algorithm For Enhanced Performance in Data Mining

Abstract:-Data mining techniques play a crucial role in extracting valuable insights from large datasets, with clustering methods being among the most widely used. The Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm is notable for its ability to identify clusters of varying shapes while effectively handling noise. However, DBSCAN faces limitations with high-dimensional data and varying density clusters, which restrict its performance in complex datasets. This thesis investigates methods to enhance the performance of DBSCAN, focusing on optimizing parameters, improving computational efficiency, and addressing density variations within clusters. We propose an advanced DBSCAN framework that integrates adaptive parameter selection and novel density-based heuristics to improve accuracy and scalability in high-dimensional data mining applications. Experimental results demonstrate that the enhanced DBSCAN algorithm achieves superior clustering accuracy, reduced computational time, and improved noise resilience compared to the traditional DBSCAN. These findings highlight the enhanced DBSCAN's potential as a robust clustering solution for real-world data mining tasks, particularly in scenarios involving large, complex datasets.

---