Egrioglu, ErolUslu, V. RezanYolcu, UfukBasaran, M. A.Hakan, Aladag C.Mehnen, JKoppen, MSaad, ATiwari, A2019-08-012019-08-012009978-3-540-89618-01867-5662https://dx.doi.org/10.1007/978-3-540-89619-7_26https://hdl.handle.net/11480/512213th World Conference on Soft Computing in Industrial Application -- 2008 -- ELECTR NETWORKWhen observations of time series are defined linguistically or do not follow the assumptions required for time series theory, the classical methods of time series analysis do not cope with fuzzy numbers and assumption violations. Therefore, forecasts are not reliable. [8], [9] gave a definition of fuzzy time series which have fuzzy observations and proposed a forecast method for it. In recent years, many researches about univariate fuzzy time series have been conducted. In [6], [5], [7], [4] and [10] bivariate fuzzy time series approaches have been proposed. In this study, a new method for high order bivariate fuzzy time series in which fuzzy relationships are determined by artificial neural networks (ANN) is proposed and the real data application of the proposed method is presented.eninfo:eu-repo/semantics/closedAccessConference Object58265+10.1007/978-3-540-89619-7_262-s2.0-79551652450N/AWOS:000269657800026N/A