A novel Shannon entropy-based backward cloud model and cloud K-means clustering

• Published in: The Journal of supercomputing 2025, Vol.81 (1), Article 65
• Main Authors: Anjali, Gupta, Anjana
• Format: Article
• Language: English
• Subjects: Algorithms; Cluster analysis; Clustering; Comparative studies; Compilers; Computer Science; Decision making; Entropy (Information theory); Information management; Information theory; Interpreters; Knowledge management; Processor Architectures; Programming Languages; Similarity; Uncertainty; Vector quantization
• ISSN: 0920-8542; 1573-0484
• DOI: 10.1007/s11227-024-06528-5

Shannon entropy is a fundamental metric for evaluating the informational content of events, valued for its robustness, versatility, and ability to capture essential aspects of information theory. Cloud models describe the transformation between qualitative and quantitative knowledge and handle uncertainty by addressing randomness and fuzziness, offering a framework for managing complex situations. Our research introduces Shannon entropy to enhance model accuracy and decision-making in cloud models, proposing the Shannon entropy-based backward cloud transformation algorithm. We illustrate the practical implementation of behavioral matching by surveying key factors individuals consider when choosing companions with similar traits. Subsequently, we develop a cloud K-means clustering algorithm to create cloud clusters that reflect individuals with similar characteristics. Further, cloud similarity measurement analysis identifies individuals within the clusters with the highest similarity. A comparative study demonstrates the new algorithm’s efficacy against traditional methods. This research offers a novel method for improving human decision-making under uncertainty.