GDC: A Hybrid Method for Hierarchical Data Grading and Classification

With the continuous expansion of big data scales, traditional single data grading and classification algorithms face challenges in handling complex data structures and achieving efficient querying. This paper proposes GDC, a hybrid method combining hierarchical clustering and decision trees. The method first builds a bottom-up data hierarchy using agglomerative hierarchical clustering, requiring no predefined cluster numbers. Decision trees are then employed to perform precise classification within each grade based on multi-dimensional features such as business value, sensitivity, and usage frequency. Simulated and real-world datasets are used in experiments, evaluated by accuracy and query efficiency improvement rate. The results show that GDC is effective and efficiency. This work advances data grading/classification methods for big data, with significant implications for query efficiency and data security in practice.