A Hybrid Neutrosophic and Machine Learning Model for Assessing Environmental Literacy in Biodiversity Conservation

Abstract

This study proposes the hybrid NEAML-BIOPASTAZA (Neutrosophic and Explainable Artificial Learning) model for Biodiversity and Legal-Ecological Assessment in Pastaza, which integrates multivariate statistical analysis, neutrosophic logic, and supervised machine learning to assess the relationship between environmental literacy and the effectiveness of the legal framework for biodiversity conservation in the Pastaza canton. Using a database of 350 observations, exploratory factor analysis was applied to validate the latent structure of the "environmental literacy" construct, considering variables such as legal knowledge, biodiversity perception, community participation, and media exposure. To manage the uncertainty inherent in social responses, a neutrosophic model was implemented, capturing the degrees of truth (T), indeterminacy (I), and falsity (F) of each perception. Finally, a Random Forest Classifier was used to predict the level of effective conservation, identifying the most relevant factors in local ecological decision-making. The combined approach allows for a more comprehensive and explanatory view of the problem, highlighting the need to strengthen environmental education, legal implementation, and community participation as pillars for the sustainable management of Amazonian biodiversity.