Diffusion-TS: 通用时间序列生成的可解释扩散
Denoising diffusion probabilistic models (DDPMs) are becoming the leading paradigm for generative models. In this paper, we propose Diffusion-TS, a novel diffusion-based framework that generates high-quality multivariate time series samples using an encoder-decoder transformer with disentangled temporal representations, aiming to satisfy both interpretability and realness.
本研究提出了一种时间序列扩散方法(TSDM),通过改进U-net架构和结合扩散模型的原理,对时间序列进行生成和故障诊断实验证明。结果显示,TSDM能够准确生成时序中的单频和多频特征,并提高小样本故障诊断准确率。