| Autor: |
Roy, Anik, Islam, Mukitul, Karim, Mehrab, Ahmed, Kazi Arman, Khan, Ashiqur Rahman, Uddin, Mezbah, Xames, Md Doulotuzzaman |
| Zdroj: |
International Journal of Information Technology; 20230101, Issue: Preprints p1-10, 10p |
| Abstrakt: |
Multi-criteria decision-making (MCDM) methods are used to deal with multiple properties of products to classify them precisely instead of traditional ABC analysis. However, to remain competitive, companies must often introduce newly manufactured products and reclassify existing inventory, which is time-consuming. To reduce time consumption and disruption, machine learning (ML) methods are employed to forecast the class of newly added inventory items. The goal of this research is to compare support vector machine (SVM), and K-nearest neighbours (KNN) with the MCDM method Technique for Order Performance by Similarity to Ideal Solution (TOPSIS) to determine the most accurate ML model for multicriteria inventory item classification. Initially, ABC analysis is used to categorize existing inventory items based on TOPSIS performance parameters, and then KNN and SVM are implemented to forecast the class of newly added inventory items. Following this, performance measures for each algorithm are calculated. From our case studies, the average training and test accuracy of the KNN model is 98.575 and 97.17% respectively. On the other hand, the average training and test accuracy of the SVM model is 69.143 and 82.5% respectively. The findings demonstrates that the KNN model had greater training and test accuracy than the SVM model. |
| Databáze: |
Supplemental Index |
| Externí odkaz: |
|
Full text from SpringerLink