Conventional displays of structures in data compared with interactive projection-based clustering (IPBC)

Abstract Clustering is an important task in knowledge discovery with the goal to identify structures of similar data points a dataset. Here, focus lies on methods that use human-in-the-loop, i.e., incorporate user decisions into clustering process through 2D and 3D displays data. Some these interactive approaches fall category visual analytics emphasize power such interactively various types datasets or verify results algorithms. This work presents new method called projection-based (IPBC). IPBC open-source parameter-free using human-in-the-loop for 2.5D display identification based user’s choice dimensionality reduction method. The approach systematically compared accessible cluster twelve benchmark one additional natural Qualitative comparison 2D, empirical evaluation identified show outperforms comparable methods. Additionally, assists identifying previously unknown domain experts application.