상황
A predictive maintenance team needs to generate synthetic time-series sensor data to supplement limited real data for training failure prediction models.
이 도면에 담긴 내용
그 이면의 의사결정을 읽어 보세요.
01Select generation approach
physics-based simulation vs. data-driven generative model
02Validate realism of synthetic data
loop back to adjust parameters if not statistically realistic
Use this diagram when designing a synthetic data pipeline for machine learning in industrial IoT or predictive maintenance contexts where real failure data is scarce.