Periodicity detection in turkish stock market
Ekim Kurtulmaz, Ravşan Aziz, Ural Ucar, Tansel Özyer, Reda Alhajj · 2018
This paper provides a periodicity detection sample of the Turkish Stock Market using data mining concepts and techniques. The extraction of periodic patterns from the time series databases is a captivating area in data mining such that it has impulse to forecast and predict the behavior of time series data in the future. Given data from on a multilevel space from different industries, we find repeating trends and frequent patterns using correlation analysis and fourier spectral evaluation. Using the projection of transformed time-series data of the feature space, we indicate long-term movements, cyclic moves, seasonal variations, and random moves. Finally, we will present a simple trend analysis for time-series forecasting the periodicity.