Footwear quality evaluation using decision tree and logistic regression models

• Main Author: Tan, Swee Choon
• Format: Thesis
• Language: English; English
• Published: 2022
• Subjects: HF5415.33 Consumer Behavior.
• Online Access: https://etd.uum.edu.my/10131/1/s824479_01.pdf; https://etd.uum.edu.my/10131/2/s824479_02.pdf; https://etd.uum.edu.my/10131/

The quality of footwear is important for manufacturers and customers. It provides a comfort protection to human foot, especially who have problem with systemic disease. However, the low state of footwear quality could lead to dissatisfaction among customers. The objectives of the study are to determine the rank factors that affect the quality of footwear using decision tree methods. Then, various types of decision trees and logistic regression model are developed to gain the best classification model for predicting footwear quality performance. Besides that, logistic regression has also been used to determine the relationship between the factors and the footwear quality performance. The data related to bubble, air trap, material problem, length out of standard, improper of mould clean, colour deviation, change model or mould, machine setting and mould setting has been observed and recorded. In six-month period, there are 1528 daily data has been collected. Based on the nine factors, the most important factors are change model or mould followed by mould setting and air trap. The analysis showed that Decision Tree with Gini algorithm (three branches) in the first method prevails against the other methods with misclassification rate of 0.1307. The model can be implemented to determine the best solution to improve the quality and performance of the footwear product.