Hybrirank: A Dual-Axis Algorithm Ranking Framework via Dataset-Based and Preference-Guided Rank Fusion

Abstract

Algorithm selection remains a challenging task due to the absence of standardized tools that support informed decision-making for identifying suitable algorithms for AI and ML tasks. It plays a crucial role in designing and deploying intelligent systems. This paper presents Hybrirank, a hybrid framework that ranks algorithms by integrating two components, dataset-based benchmarking module that evaluates performance on varied domain-specific datasets using standard metrics, and a user preference modeling module that captures user-defined priorities through structured input. The system supports clustering, classification, and regression tasks, and allows users to generate rankings based on dataset performance, user preferences, or both. Each mode produces a context-aware ranking tailored to its input. In hybrid mode, HybrirankFramework leverages its core fusion mechanism to integrate data-driven performance and user priorities to generate a more selective and user-aligned recommendation. This ranking serves as a decision-support system, enabling users to select the most appropriate algorithm. Hybrirank offers a dual-layered evaluation process that enhances recommendation relevance, bridges the gap between empirical performance and practical needs, and it supports optimal selection guided by both data and priorities.