• Density-based clustering
• Data stream
• Adaptive parameters

This thesis work concerns the study of an adaptive fuzzy density-based clustering algorithm for data streams. Unlike the static case, the streaming scenario is characterized by a continuous and potentially infinite flow of incoming data, produced nowadays by a myriad of software applications and hardware platforms.
A crucial aspect of the streaming setting is the concept drift that consists of a change in the distribution of the data over time. Starting from an existing implementation, which also takes advantage of fuzziness to more appropriately model cluster borders, we develop an approach that can manage concept drift by adapting clustering algorithm parameters to the evolution of the data stream. Efficiency and effectiveness of the proposed approach are evaluated on synthetic and on high-dimensional real-world data streams.