fuzzy time series forecasting (FTSF) is a typical forecasting method with
wide application. Traditional FTSF is regarded as an expert system which leads
to lose the ability to recognize undefined feature. The mentioned is main
reason of poor forecasting with FTSF. To solve the problem,
LTSF-DNODE is proposed as a solution to the limitations of Linear-based LTSF models and Transformer-based approaches, showing better performance on various real-world datasets and exploring the impacts of regularization in the NODE framework for each dataset.