Fabric and Production Defect Detection in the Apparel Industry Using Data Mining Algorithms

Abstract

Nowadays, technology plays a crucial role in fabric production in the textile industry. The demand for high-quality products and rapidly changing economic conditions increase the significance of readymade clothing manufacturers to produce the right quality product. In addition, in order to minimize production errors, to improve and maintain process performance, it is important to identify the sources of variability during manufacturing. The defective fabric is the main reason which is causing harm to the textile business. Therefore, the proper identification of manufacturing defects leads to a successful business. When it comes to extracting meaningful insights from data and knowledge discovery, data mining has proven its significance in various fields such as business, health, finance, and education. As in all other sectors, data mining is widely used in the textile sector too. In this study, it was aimed to determine the main causes of the error with defective production data of a company that produces clothing by using data mining methods. Decision Tree, Naive Bayes, Random Forest, and Gradient Boosted Trees Algorithms were used in the research. Accuracy rate and Cohen's kappa statistics were taken for comparison algorithms in the study. While determining the main reasons for defective products, the factors of which type of products the company produces for its customers, the sizes of the defective products, types of defects, and explanations were taken into consideration. The most common mistakes in sewing production and the main source of the error were evaluated. According to the results, suggestions were made for the company to take various measures.

Developing a data-driven operational guide for the texturized yarn production process: data mining and intelligence approach

“…(2020b) employed AutoML to predict tear strength in both the warp and weft directions. (Ersöz et al ., 2021) employed various methods, including Naïve Bayes, gradient boosted trees, random forest and decision trees, to identify the primary causes of defects in a clothing manufacturing company. In another study, an adaptive neuro-fuzzy inference system and image processing techniques were utilized to determine the wrinkle grade of fabric (Hesarian et al ., 2020).…”

Artificial-Neural-Network-Based Predicted Model for Seam Strength of Five-Pocket Denim Jeans: A Review

“…The occurrence of defects was determined according to the defect type, model number, production size, and product type. The accuracy rates of the model were compared, and it was seen that "decision tree" algorithms had higher accuracy rates than other classifier algorithms that were used [87].…”