Use of Fuzzy Time Series to Generate Linguistic Descriptions of Noise Pollution

In this paper we present a proposal to represent sets of Time Series as a Fuzzy time Series (FTS) composed of linguistic labels previously defined in a set. In the definition of this set of linguistic labels the knowledge from an expert in the domain of application can be Incorporated. First, the input set of TS is represented by means of FTS whose values are fuzzy sets. Then, these fuzzy sets are transformed into linguistic labels taken from the predefined set and then a new FTS is obtained whose main difference with the previous one is that it is more interpretable. The Monte Carlo method has been used to perform this transformation. Finally, to show the validity of our proposal, in the experimentation a case use in the generation of linguistic descriptions of noise pollution is presented.