A New Attribute Reduction Algorithm Under Tolerance-Based Relation for Rough Neutrosophic Decision System

Abstract

Rough set has been successfully combined with other mathematical frameworks to improve attribute reduction. In particular, attribute reduction is essential for processing and analyzing the Rough Neutrosophic Decision System. The rough neutrosophic set provides an effective framework for managing vagueness, inconsistency and incomplete information. This hybrid model allows a more flexible representation of real-world data by incorporating truth, indeterminacy and falsity membership functions. This study introduces a novel attribute reduction technique based on tolerance relations in the context of rough neutrosophic sets, employing rough neutrosophic numbers to express information values. The proposed method includes the formation of lower and upper approximations and defines the degree of dependency between decision attributes and conditional attributes. An algorithm is developed to implement the approach  and a detailed example in coastal erosion is provided to validate its practical application. Experimental outcomes demonstrate that the method efficiently identifies relevant and non-relevant attributes, thereby enhancing the decision-making process. The proposed method not only improves the precision of data but also strengthens the robustness of intelligent systems when dealing with complex and uncertain datasets.