Prediction of production line performance using neural networks

Abstract

The use of artificial neural networks in many fields is still on the increase. The paper deals with application of neural networks as a data mining method to a prediction of the production line performance. Performance of production line was defined by output indicators like number of finished products, flow time and work in progress production. Predictive model was implemented in the program STATISTICA Data Miner, therefore this paper brings also short overview of used options. The overall quality of learned networks was evaluated. PMML file was created for fast deployment to new data and better decision making. Neural networks provide an effective analyzing and diagnosing tool to understand and simulate the behavior of the plant, and can be used as a valuable performance assessment tool for decision makers.