Time series: Advanced methods of analysis and forecast
2014 IEEE World Congress on Computational Intelligence (WCCI 2014)
Beijing, China, July 6-11, 2014

Aims and scope
For a long time, time series processing was the prerogative of
statistics. This tradition was broken with the emergence of such
modern methods of data analysis as neural networks, fuzzy sets, rough
sets, etc. These non-statistical methods are especially useful in the
cases when time series are short and/or cannot be characterized as
stochastic processes, i.e., time series observations are vague and do
not show regular
behavior. Typical examples are financial and medical time series. The
aim of this special session is to present recent developments and
trends in time series analysis and forecasting including fuzzy time
series and granular time series. We are aware that representation and
processing of time series are tightly connected. We invite
contributions that extend traditional ways of modeling time series and
propose adequate methods for their processing. A special focus will be
made on processing and forecasting of multi-variate time series.

We expect that this special session will collect fundamental and
theoretical results, show other than the approaches mentioned above
and overview the existing ones. We invite contributions focused on
(but not limited to) the following topics:

– time series trends, tendencies and their models,
– fuzzy time series, granular time series and their models,
– time series classification and forecast,
– forecasting methods: regression, fuzzy regression, fuzzy/linguistic
IF-THEN rules,
– forecasting of multi-variate time series,
– F-transform and fuzzy natural logic in time series analysis and forecast,
– special classes of time series.