<p>An emerging field within the internet of things (IoT) known as cognitive IoT (CIoT) has gained attention due to the growing interest in integrating cognitive capabilities into IoT systems. Similar to IoT, which consists of a network of interconnected devices, CIoT generates vast amounts of data that are diverse, unpredictable, and time-sensitive. Consequently, efficiently extracting computational knowledge from these large volumes of real-time sensor data presents a significant challenge for CIoT. This study proposes a novel approach to identifying meaningful concepts and patterns through imprecise reasoning. Initially, the proposed method addresses the impact of noisy inputs, followed by the construction of a concept lattice. The weight of the generated concepts is computed using plausibility values, allowing for the selection of minimally weighted (significant) concepts. In the next phase, relevant data are extracted from these significant concepts, which are then combined to form meaningful patterns. The effectiveness of the proposed approach is empirically evaluated using weather data spanning 21.25&#xa0;years. A variety of metrics were employed to validate the proposed method, and the results, with an accuracy exceeding 99.57%, demonstrate that the proposed algorithm outperforms existing competitors.</p>