Data driven solution to predictive maintenance

This thesis proposes solutions to predictive maintenance in the context of industries that are in transition to industry 4.0. Several data-driven methods have been tried out and the results are compared and evaluated. As a result, we conclude that predictive maintenance using data-driven methods in industries are feasible. However, the availability of data and the selections of meaningful data are critical to the project outcome. Furthermore, the thesis has presented the challenges arises during the process due to the difference in theoretical and practical settings.