Enhanced Clustering Framework for Unveiling Hidden Patterns in IoT-Based Structured and Unstructured Data

Clustering algorithms play a pivotal role in uncovering hidden patterns and relationships within diverse datasets. This study discusses about an enhanced real-time clustering approach based on the Birch methodology, optimized for analyzing both structured and unstructured data. Structured data, consisting of IoT sensor readings, and unstructured image data, captured using Raspberry Pi devices deployed in various locations, were analyzed to identify meaningful patterns. For structured data, the algorithm effectively clustered environmental parameters, revealing correlations and anomalies. For unstructured data, high-dimensional image features extracted using a pre-trained VGG16 model were clustered, uncovering patterns in texture, color, and structure. Experimental results demonstrate the algorithm’s scalability and efficiency, achieving high-quality clusters with average Silhouette Scores of 0.87 and 0.81 for structured and unstructured datasets, respectively. These findings highlight the algorithm’s adaptability and its potential for real-time analytics in diverse applications such as environmental monitoring and image-based decision-making systems.