Enhancing Supply Chain Risk Identification: Analyzing the Impact of LLM Parameters for precise Classification
- This study investigates the impact of Large Language Model (LLM) parameters, specifically temperature and top P, on Supply Chain Risk Detection (SCRD). With a heightened focus on Supply Chain Risk Management (SCRM) using AI, the research employs a Design of Experiments (DoE) approach. The results reveal optimal temperature values for valid assessments in SCRD applications. The study emphasizes the importance of tailored LLM parameter settings, contributing insights for future research and practical applications in enhancing supply chain resilience. Suggestions for incorporating Response Surface Methodology (RSM) and refining the process are proposed for further investigation.
| Author: | Linus KühlORCiD, Jost WiethölterORCiD, Michael Dircksen |
|---|---|
| URL: | https://www.islconf.org/wp-content/uploads/2024/07/ISL_2024_Proceedings_Final.pdf |
| ISBN: | 978-0-85358-354-7 |
| Parent Title (English): | Building sustainable connectivity through logistics and supply chains : proceedings of the 28th International Symposium on Logistics (ISL 2024) : 07-10th July 2024 |
| Publisher: | Centre for Concurrent Enterprise, Nottingham University Business School |
| Place of publication: | Nottingham, UK |
| Document Type: | Conference Proceeding |
| Language: | English |
| Date of Publication (online): | 2024/07/12 |
| Date of first Publication: | 2024/07/08 |
| Provider of the Publication Server: | FH Münster - University of Applied Sciences |
| Release Date: | 2024/07/22 |
| Tag: | Artificial Intelligence; Data Analytics; Design of Experiments; Large Language Model; Logistics |
| Pagenumber: | 9 |
| First Page: | 197 |
| Last Page: | 205 |
| Faculties: | Wirtschaft (MSB) |
| Publication list: | Dircksen, Michael |
| Wiethölter, Jost | |
| Licence (German): |  Bibliographische Daten |