Probabilistic Density-based Adaptive Clustering for Streaming Data

The Internet of Things(IoT) has been applied in many real-world scenarios to provide intelligent solutions in recent years. The biggest challenge in these applications is the enormous volume of data generated from the devices. Adequate mechanisms are demanded to process this data in real-time to generate valuable information. In many IoT scenarios, a grouping of generated data is effective in deriving insights. Clustering techniques are beneficial for grouping data when the label is partially or fully not available. This paper proposes a probabilistic clustering algorithm for IoT data streams, where the number of clusters need not be defined in advance. In addition, the clusters are managed adaptively using a maintenance system. The model has experimented with two publicly available IoT data sets: weather and pollution data. We compared cluster quality with existing algorithms, and the results demonstrate the effectiveness of the proposed clustering scheme.