CRFormer: Complementary Reliability Perspective Transformer for Automotive Components Reliability Prediction Based on Claim Data

Reliability prediction has been studied in many industries for managing stocks and reducing quality assurance costs production costs. Particularly, the automotive industry, reliability is performed based on two automobile perspectives, time mileage. To maximize cost savings, researchers attempted with short-term inputs. However, limited information inputs resulted unsatisfactory results long warranty periods. Additionally, overall evaluation metrics could not reflect pattern-wise performance, such as increasing failure patterns. This study proposes Complementary perspective Transformer (CRFormer) encoder to achieve enriched representations from a input sequence. CRFormer fuses different features compensate input. The performance of evaluated claim data accumulated over 16 years. Results showed that compared previous methods terms overall, pattern-wise, pattern similarity metrics, achieved outstanding mileage prediction. Lastly, visualization survival analysis accurate model can be used support decision-making reduce