A Fault Forecasting Approach Using Two-Dimensional Optimization (TDO)
Primary Area: optimization
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Keywords: Fault Detection; Tuple Selection; Feature Selection
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
TL;DR: Two-Dimensional Optimization (TDO)
Abstract: Data preparation plays a pivotal role in every machine learning-based approach, and this holds true for the task of detecting claims in the automotive industry as well. Handling high-dimensional feature spaces, especially when dealing with imbalanced data, poses a significant challenge in sectors where a vast amount of data accumulates over time. Machine learning models trained on highly imbalanced data often result in unreliable and untrustworthy predictions. Therefore, addressing the aforementioned issues is essential during the data pre-processing phase. In this paper, we propose an innovative two-dimensional optimization approach to effectively address the challenge of highly imbalanced data in the context of fault detection. We employ a heuristic optimization algorithm called Genetic Algorithm to concurrently reduce both the data point tuples and the feature space. Furthermore, we constructed and evaluated two-dimensional reduction using particle swarm optimization (PSO) and Whale optimization algorithms. The empirical results of the proposed techniques on the data collected from thousands of vehicles show promise.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 3395
Loading