Time series; fuzzy rules; ensembles; Fuzzy Rule Based Ensemble; fuzzy GUHA; linguistic associations; perception-based logical deduction

There are many various methods to forecast time series. However, there is no single forecasting method that generally outperforms any other. Consequently, there always exists a danger of choosing a method that is inappropriate for a given time series. To overcome such a problem, distinct ensemble techniques are being proposed. These techniques combine more individual forecasting methods. In this contribution, we employ the so called fuzzy rule-based ensemble to determine the weights based on time series features such as trend, seasonality or stationarity. For identification of fuzzy rule base, we use linguistic association mining. An exhaustive experimental justification is provided.