Abstract
We have various kinds of time series data such as a temperature record. We understand them via their linguistic expressions in a natural language. We have proposed a method to extract their linguistic expressions of time series with a global trend and some local features, while Kacprzyk and Wilbik proposed several methods of linguistic summarization. In our method, we have three aggregated values with dividing a whole period into three terms by fuzzy sets. We adopt the weighted average of time series values by the membership values of the fuzzy sets for the aggregated values. In this paper, we adopt the orderd weighted average (OWA) operator by Yager, which is widely used in decision-making. It can implement various aggregation operators by setting different weights such as the maximum operator, the minimum operator and the mean operator. We, therefore, have better linguistic expressions for global trend without influence of extreme values in local features.
