Neutrosophic OWA-TOPSIS Model for Decision-Making in AI Systems with Large Volumes of Data
AI systems require transformations of big data critical to processing because the findings are based upon incomplete, inconsistent, and/or biased findings which mean that findings and subsequent achieved expectations will inevitably have limitations. This is problematic when engaging multi-criteria...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
University of New Mexico
2025-05-01
|
| Series: | Neutrosophic Sets and Systems |
| Subjects: | |
| Online Access: | https://fs.unm.edu/NSS/49.%20NeutrosophicOWA-TOPSISBigData.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | AI systems require transformations of big data critical to processing because the findings are based upon incomplete, inconsistent, and/or biased findings which mean that findings and subsequent achieved expectations will inevitably have limitations. This is problematic when engaging multi-criteria decision making with big datadriven projects like infiltration of personalized suggestions, AI-based medical diagnostics, and risk reduction efforts where any decision making with deficient data can reduce effective capabilities. The literature suggests that TOPSIS and OWA operators enable the prioritization of alternatives given ranked data; however, there is a gap in the literature regarding the suitability of decision making techniques to prioritize plithogenic uncertainty. Yet this is relevant because in life, things/ideas/situations aren't true or false—they're somewhere indeterminate. Thus, this paper presents a new, hybrid approach that combines OWA-TOPSIS with neutrosophic sets to determine how much truth, falsity, and indeterminacy exist for specific criteria within the decision making process. By adjusting neutrosophic distances and executing an entropy-dependent OWA weight to create a final decision ranking within the presented case, data can accurately render situations where customer reviews for products have good and bad features or scenarios where machine learning algorithm effectiveness has sometimes opposing results. This case study's findings indicate that this hybrid idea is more accurate than traditional TOPSIS, 89.4% vs. 82.1%, and more stable even at high uncertainty levels. The theoretical contribution to the academic literature expands notions of AI multi-criteria decision making process; the practical application lends itself to scalable possibilities within big data reliant cases, especially predictive sentiment analysis or resource allocation/optimization. The feasibility of neutrosophic applications within distributed interfaces (like Spark) shows the promise for real-time applications explored without delay. |
|---|---|
| ISSN: | 2331-6055 2331-608X |