HDBMS: A Context-Aware Hybrid Graph Traversal Algorithm for Efficient Information Discovery in Social Networks

TL;DR

HDBMS is a novel context-aware hybrid graph traversal algorithm that adaptively balances depth and breadth exploration using probabilistic node relevance, improving efficiency and effectiveness in social network information discovery.

Contribution

The paper introduces HDBMS, a new adaptive graph traversal algorithm that dynamically adjusts exploration strategies based on probabilistic relevance estimates, outperforming traditional methods.

Findings

HDBMS outperforms traditional algorithms in identifying meaningful paths.

HDBMS maintains competitive computational efficiency.

HDBMS is effective in social network analysis and information retrieval.

Abstract

Graph-searching algorithms play a crucial role in various computational domains, enabling efficient exploration and pathfinding in structured data. Traditional approaches, such as Depth-First Search (DFS) and Breadth-First Search (BFS), follow rigid traversal patterns -- DFS explores branches exhaustively, while BFS expands level by level. In this paper, we propose the Hybrid Depth-Breadth Meaningful Search (HDBMS) algorithm, a novel graph traversal method that dynamically adapts its exploration strategy based on probabilistic node transitions. Unlike conventional methods, HDBMS prioritizes traversal paths by estimating the likelihood that a node contains the desired information, ensuring a more contextually relevant search. Through extensive experimentation on diverse directed graphs with varying structural properties, we demonstrate that HDBMS not only maintains competitive computational efficiency but also outperforms traditional algorithms in identifying meaningful paths. By integrating probabilistic decision-making, HDBMS constructs an adaptive and structured traversal order that balances exploration across depth and breadth, making it particularly effective in applications such as information retrieval, social network analysis, and recommendation systems. Our results highlight the robustness of HDBMS in scenarios where the most valuable connections emerge unpredictably, positioning it as a powerful alternative to traditional graph-searching techniques.