Cost-sensitive feature selection for multi-label classification: multi-criteria decision-making approach

Applied Computing and Informatics, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 5, 2025

Purpose In multi-label classification, selecting the most relevant features is crucial for enhancing predictive performance and reducing computational complexity. Real-world scenarios often involve significant costs in data acquisition, including time, financial resources. However, existing feature selection methods overlook associated costs. Design/methodology/approach Multicriteria decision-making (MCDM) has emerged as a powerful tool addressing complex problems involving multiple, conflicting criteria. This study proposes novel cost-sensitive method that fuses importance with cost within an MCDM framework. The proposed transforms problem into by leveraging mutual information. Furthermore, were converted Fermatean fuzzy sets, simple weighted sum product (WISP) was employed to rank based on their relevance labels Findings Extensive experiments conducted ten benchmark datasets against five evaluation metrics demonstrated superiority of while minimizing consistently outperforming methods. Originality/value Unlike integrate through penalties select via greedy search, approach adopts MCDM-based strategy ranking. aims achieve globally optimal outcomes balancing trade-offs between objectives, marking advancement over techniques. Graphicalabstract

Language: Английский

A Fuzzy Multi-Criteria Approach for Selecting Open-Source ERP Systems in SMEs Using Fuzzy AHP and TOPSIS

Jurnal Optimasi Sistem Industri, Journal Year: 2025, Volume and Issue: 23(2), P. 167 - 187

Published: Jan. 31, 2025

In a rapidly growing and competitive business era, selecting an open-source Enterprise Resource Planning (ERP) system is critical step to support the efficiency effectiveness of company operations. This research aims propose innovative methodology by integrating fuzzy Analytical Hierarchy Process (fuzzy AHP) Technique for Order Preference Similarity Ideal Solution TOPSIS) improve ERP selection process. The method involves eight criteria 26 sub-criteria comprehensively evaluate 11 alternatives, specifically SMEs in transportation services sector Indonesia. System quality has been identified as factor system, with particular emphasis on aspects such security reliability. These are considered most influential determining suitability system. analysis further indicates that 10th alternative best choice, consistently outperforming others meeting defined criteria. Additionally, sensitivity confirmed robustness this demonstrating its stability despite changes weights. Beyond practical implications SMEs, contributes versatile evaluation framework can be adapted other industries seeking effective solutions. findings emphasize importance structured decision-making technology adoption, offering comprehensive reliable guidance organizations aiming optimize their operations through systems. study not only bridges gap but also establishes methodological foundation future applications across diverse industry sectors.

Language: Английский