Enhancing decision analysis with a large language model: pyDecision a comprehensive library of MCDA methods in Python

Abstract

Purpose Multicriteria decision analysis (MCDA) has become increasingly essential for decision-making in complex environments. In response to this need, the pyDecision library, implemented in Python and available at https://bit.ly/3tLFGtH, has been developed to provide a comprehensive and accessible collection of MCDA methods. The purpose of this paper is to demonstrate the capabilities of pyDecision and illustrate how it can be effectively applied to solve multicriteria decision problems. Design/methodology/approach The pyDecision offers 70 MCDA methods, including analytic hierarchy process, technique for order of preference by similarity to ideal solution, the preference ranking organization method for enrichment evaluation and ELimination Et Choix Traduisant laRÉalité families. Beyond offering a vast range of techniques, the library provides visualization tools for more intuitive results interpretation. In addition to these features, pyDecision has integrated ChatGPT, an advanced large language model, where decision-makers can use ChatGPT to discuss and compare the outcomes of different methods, providing a more interactive and intuitive understanding of the solutions. Findings Large language models are undeniably potent but can sometimes be a double-edged sword. Its answers may be misleading without rigorous verification of its outputs, especially for researchers lacking deep domain expertise. It is imperative to approach its insights with a discerning eye and a solid foundation in the relevant field. Originality/value With the integration of MCDA methods and ChatGPT, pyDecision is a significant contribution to the scientific community, as it is an invaluable resource for researchers, practitioners and decision-makers navigating complex decision-making problems and seeking the most appropriate solutions based on MCDA methods.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 26 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.