Conscious points and patterns extraction: a high-performance computing model for knowledge discovery in cognitive IoT

Incorporating cognition into the design and architecture of the Internet of Things (IoT) has recently been the subject of much research, giving rise to a new subfield known as the cognitive IoT (CIoT). Consequently, CIoT takes on some characteristics and challenges of the IoT. Many applications in CIoT generate a massive amount of data that require meaningful insight in a computationally efficient manner. Therefore, this research proposed a novel way to extract meaningful insight from huge amounts of data using probabilistic values. Along with that, it finds the most informative observation from each cluster based on the amount of information associated with each sensory observation. Subsequently, the modified Lagrangian L1 algorithm is used to extract conscious points, which are further combined, and the highest entropy row is selected as the conscious pattern. Experimenting with 21.25 years of environmental data and cross-validating using many metrics, it outperforms current approaches.