Time Series Mining by Fuzzy Natural Logic and F-Transform

Author

Novak, Vilem ; Perfilieva, Irina

Author_Institution

Inst. for Res. & Applic. of Fuzzy Modeling NSC IT4Innovations, Univ. of Ostrava, Ostrava, Czech Republic

fYear

2015

fDate

5-8 Jan. 2015

Firstpage

1493

Lastpage

1502

Abstract

In this paper, we discuss application of special soft computing methods, namely the fuzzy natural logic and fuzzy transform, to the problem of mining information from time series. The mined information is formulated in natural language. We discuss the following applications: reduction of size of time series, extraction of its trend-cycle, linguistic characterization of its future course and forecast of future values, finding perception ally important points, extraction and linguistic evaluation of local trend in a given area, and summarization of characteristics of time series using generalized intermediate quantifiers.

Keywords

computational linguistics; data mining; fuzzy logic; fuzzy set theory; natural language processing; time series; transforms; F-transform; fuzzy natural logic; fuzzy transform; generalized intermediate quantifiers; information mining; linguistic characterization; linguistic evaluation; local trend extraction; natural language; size reduction; soft computing methods; time series mining; Bismuth; Context; Fuzzy sets; Natural languages; Pragmatics; Time series analysis; Transforms; F-transform; Fuzzy Natural Logic; Fuzzy transform; Information mining; Time series;

fLanguage

English

Publisher

ieee

Conference_Titel

System Sciences (HICSS), 2015 48th Hawaii International Conference on

Conference_Location

Kauai, HI

ISSN

1530-1605

Type

conf

DOI

10.1109/HICSS.2015.181

Filename

7069990