Robust parameter design of products and processes with an ordinal categorical response using random forests

Thesis Type: Postgraduate

Institution Of The Thesis: Orta Doğu Teknik Üniversitesi, Faculty of Engineering, Department of Industrial Engineering, Turkey

Approval Date: 2018




In industrial organizations, manufacturers aim to achieve target product performance with minimum variation. For that reason, finding optimal settings of product and process design parameters that make it possible to consistently achieve target product performance is an important design problem. In this study, we propose an alternative method to solve this problem for the case of an ordinal categorical product/process response. The method utilizes Random Forest (RF) for modelling mean and variance of the response at a given set of parameter settings. The method uses different weighting strategies of Random Forest, and it is applied on three case problems. Two of the case problems are of the larger-the-better type, and the other one is of the smaller-the-better type. In addition, obtained results are compared with those of previous studies that used the same data sets. In comparing the results, classification performance, probability of observing target class, and both location and dispersion of results are considered. Advantages and disadvantages of the proposed method are discussed.