Comparing approaches to equipment-failure classification
This academic project examined neural-network, decision-tree, and logical-analysis models for five industrial equipment-failure modes. My work focused on data preparation and comparison of the classification approaches.
The workflow
Feature engineering and SMOTE were part of the preparation workflow. The model comparison considered different ways to use equipment data to distinguish failure modes.
The project extended my work on maintenance systems into data analysis: here, the engineering task was preparing the inputs and comparing classification methods for equipment-failure data.
Project context
The available project record supports the methods and scope summarized here. An evaluation report, dataset, and code are not publicly linked, so this page does not report a predictive-accuracy figure.