Applying data mining for manufacturing cycle time prediction : A case study Scancom Vietnam Limited Company

Abstract

Late delivery is considered as a big problem in Scancom Vietnam Limited Company. One of reasons effect on late delivery is that the production planning and the actual production are not consistent. The aim’s research is minimizing difference of cycle time between planning and manufacturing in real. Data mining is applied to estimating reliable manufacturing cycle time for a new product in order to make right decision in negotiation on due dates of multiple orders between manufacturers and customers. A predictive model is built based on historical data using regression approach. The actual data which collected from Cutting and Bending sector at Aluminum Factory in Scancom Vietnam Limited Company is used to train and test. The estimation performance of regression approach is compared with results that executed by neural network approach and k-nearest-neighbor approach in data mining. Applying this method to identify the cycle time is more accurate that improves production units in planning, monitoring and ensuring efficient utilization of manufacturing capacity. Keywords: cycle time prediction, data mining, linear regression, neural network, k- nearest neighbor.