NeutroLab: A Unified Library for Neutrosophic Computing

Abstract

This paper introduces NeutroLab, a comprehensive open-source Python library for neutrosophic computing, available on PyPI. NeutroLab unifies four major research contributions: (1) Five neutrosophication methods for transforming crisp data into neutrosophic triplets <T, I, F>, including a novel K-Means approach that achieves true component independence; (2) N-fsQCA v2.0, a neutrosophic extension of fuzzy-set Qualitative Comparative Analysis (fsQCA) with variance-based indeterminacy and 11 causal archetypes; (3) NML (Neutrosophic Meta-Learning) for post-hoc uncertainty analysis of Tsetlin Machines with a computational overhead of ~4%; and (4) IFAO (Indeterminacy-First Aggregation Operator), a novel MCDM paradigm where indeterminacy is the ontological foundation from which truth and falsity emerge. The library provides a unified API, extensive documentation, and is available via PyPI (pip install neutrolab). Experimental validation demonstrates superior uncertainty quantification compared to existing approaches across multiple domains, including medical diagnosis, qualitative research, explainable AI, and multi-criteria decision-making.

Keywords