作者: Cong Tran , Won-Yong Shin , Sang-Il Choi
DOI: 10.1109/ACCESS.2018.2845843
关键词:
摘要: An imprecise region is referred to as a geographical area without clearly defined boundary in the literature. Previous clustering-based approaches exploit spatial information find such regions. However, prior studies suffer from following two problems: subjectivity selecting clustering parameters and inclusion of large portion undesirable (i.e., number noise points). To overcome these problems, we present DIR-ST 2, novel framework for delineating an by iteratively performing density-based applications with (DBSCAN) along not only spatio–textual but also temporal on social media. Specifically, aim at finding proper radius circle used iterative DBSCAN process gradually reducing each iteration which acquired all resulting clusters leveraged. Then, propose efficient automated algorithm via hierarchical clustering. Experimental results show that virtue significant reduction region, our 2 method outperforms state-of-the-art approach employing one-class support vector machine terms $\mathcal {F}_{1}$ score comparison precisely regions regarded ground truth, returns apparently better delineation The computational complexity analytically numerically shown.