Date

2022-8

Author

Teköz, Begüm

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

323
views

520
downloads

Cite This

Forecasting claim rate under warranty allows companies to optimize their production processes, reduce warranty costs and maintain customer satisfaction. In the case of a production crisis, the poor performance of the claim rate forecast negatively affects business processes. This thesis aims to improve the business processes of many departments, including production, research and development, quality, and after-sales, by forecasting the number of monthly claims in each service group. In this study, warranty data obtained from an automotive industry is used to forecast three months data for twenty-five different in-service warranty performance groups using statistical and machine learning algorithms. Specifically, statistical approaches including ARIMA, TBATS and ETS models and machine learning methods including random forest, support vector regression, XGBoosting, feed forward neural network, long short-term memory neural network, and Bayesian regularized neural network are employed. The performance of the models is compared with the Wilcoxon signed-rank test, and the results show that the best performing models are machine learning methods and the random forest model.

Subject Keywords

Time series analysis, Machine Learning, Warranty performance

URI

https://hdl.handle.net/11511/99518

Collections

Graduate School of Natural and Applied Sciences, Thesis