A Neural Network Approach to Predict Duration in Conformity for Predictive Manufacturing

Currently, expert knowledge is sometimes the only way to predict the duration of the manufacturing process in relation to conformity. While expert knowledge is difficult to come by and often limited to experience. This paper aims to predict the duration of the manufacturing process by utilizing machine learning and big data. This paper utilizes a real industrial case with the use of neural network regression model that aims to expand the body of knowledge available in the area of predictive manufacturing research. The result of the model suggests the neural network regression model can result in a feasible outcome, while in the meantime overfitting did occur. Mean-squared error, relative squared error, relative absolute error and the correlation coefficient was computed. The paper also addresses the limitations and limited scope of findings and makes suggesting moving forwards.