Pythagorean Neutrosophic Normal Interval-Valued Weighted Averaging Approach for Sustainable Financial Risk Prediction

Neutrosophic set (NS) is a prevailing logic aimed at facilitating the understanding of inconsistent and indeterminate data; several kinds of complete or incomplete data can be described as interval-valued NS (IVNS). This study presents aggregation operator for IVNSs and prolongs the generalized weighted aggregation (GWA) operations to congruently work with IVNS information. Also, these results are formulated as IVNSs that are represented by indeterminate, truth, and false degrees. The tremendous growth of financial innovation offers a several convenience to people’s lives and production and brings many security risks to financial technology. To avoid financial risk, an improved way is to construct an accurate warning mechanism before the financial risk takes place, not to solve this matter after the risk outbreak. Recently, deep learning (DL) has delivered outstanding results in the natural language processing and image recognition areas. Thus, researcher used DL techniques for the financial risk prediction and obtained satisfactory results. This study develops a new Pythagorean Neutrosophic Normal Interval-Valued Weighted Averaging for Financial Risk Prediction (PNNIVWA-FRP) method using sustainable development. The objective of the PNNIVWA-FRP method is to have two dissimilar stages of processes. Initially, financial data are classified by the PNSNIVWA technique. This method is used for its highest proficiency in managing imprecision and uncertainty in financial data, containing incomplete and ambiguous data. Second, the classified parameter is fine-tuned by means of Glowworm Swarm Optimization (GSO) technique. Based on the luminescent communication of glowworms, GSO is proficient at navigating multidimensional, complex search spaces for identifying better solutions. The empirical findings on benchmark dataset demonstrate the effectiveness of the PNNIVWA-FRP method, showcasing significant development in prediction results than classical approaches.