Production forecasting and quality classification with machine learning techniques: Application in a textile plant

Artificial intelligence, one of the pioneers of technological developments, touches every point of our lives and continues to provide convenience to our lives. The concept of artificial intelligence, which takes place in many areas such as health, security, transport and financial services in our lives, is also of great importance in businesses. Artificial intelligence, which takes place in various fields of businesses, is more widely used especially in companies in the production process. With the digital transformation brought by Industry 4.0, the use of artificial intelligence-supported systems in production processes is becoming widespread and provides great advantages to businesses. As in every sector, artificial intelligence has advantages in the textile industry. Among these advantages, increasing efficiency in production processes, reducing costs and improving product quality stand out. In this study, the production and quality processes of a sock company operating in the textile sector are analysed. The study consists of two stages. In the first stage, production data for the years 2022 and 2023 and production forecasts for the first three months of 2024 were made using machine learning regression algorithms, a sub-branch of artificial intelligence. In the second stage, the defect types and product characteristics that emerged in the production in July 2023 were quality classified using machine learning classification algorithms. Defect types were scored by the quality control unit in the enterprise. In this application, Python programming language was used and the coding process was performed in Spyder interface. When the results are analysed, it is determined that the linear regression model is suitable for predicting the production numbers for the enterprise. In addition, it was observed that the decision tree model gave more effective results than other models in determining quality classes. This study emphasises the importance of using machine learning and artificial intelligence in production processes and also shows that analyses can be easily performed on uncertainties that may occur in any field. At the same time, it is predicted that the analysis will make a significant contribution to the literature in the fields of artificial intelligence and machine learning.