sci-FTS: Using soft clustering on Intrinsic Mode Functions to model Fuzzy Time Series
Marcos Vinícius dos Santos Ferreira, Ricardo A. Rios, Tatiane Nogueira Rios · Software Impacts · 2022
This manuscript introduces a new software, sci -FTS, which models time series by combining Signal Processing tools and Fuzzy Set Theory. Firstly, time series are decomposed into Intrinsic Mode Functions (IMFs), emphasizing instantaneous frequencies and amplitudes. Secondly, sci -FTS combines IMFs to extract deterministic influences, removing noises. Next, sci -FTS executes an algorithm that finds an adequate space partitioning to produce the fuzzy sets. Finally, Fuzzy Time Series steps are considered to predict observations. Our contributions are twofold: sci -FTS finds out similar patterns in observations to better model the universe of discourse; ii) models produced by sci -FTS overcome studies from the literature.