Services on Demand
Journal
Article
Indicators
- Cited by SciELO
- Access statistics
Related links
- Similars in SciELO
Share
Contaduría y administración
Print version ISSN 0186-1042
Abstract
MEDINA REYES, José Eduardo; CRUZ AKE, Salvador and CABRERA LLANOS, Agustín Ignacio. New hybrid fuzzy time series model: Forecasting the foreign exchange market. Contad. Adm [online]. 2021, vol.66, n.3, 00008. Epub Feb 07, 2022. ISSN 0186-1042. https://doi.org/10.22201/fca.24488410e.2021.2623.
This work develops a comparison between the volatility prediction of traditional time series models (ARIMA, EGARCH and PARCH), against two new proposed models based on fuzzy theory (FTS-Fuzzy ARIMA Tseng's and FTS-Fuzzy ARIMA Tanaka's). To make this comparison, we estimated the Mexican peso - US dollar exchange rate yield from January 2008 to December 2017. Our main result is that the models based on fuzzy theory generate a better estimate of the volatility. The fuzzy models show a smaller least forecast error than the traditional time series in both; in and out of sample tests; for the volatility in the yield of the Mexican peso - US dollar exchange rate. Therefore, the fuzzy models showed higher efficiency and better reflects the market information.
Keywords : Fuzzy logic; Fuzzy ARIMA, Fuzzy time series; Fuzzy linear regression.