Association Rules Mining from Time Series Based on Rough Set

A method of mining association rules from time series based on rough set is introduced. To clean the data, Fourier transformation is employed, and LPF operator is adopted. Partial and overall features of a time series are defined, and some innovative methods for extracting features from a time series or for segmenting a time series are proposed. Thereafter, a discretization technique that will produce symbols with equiprobability is adopted to discretize the features since rough set can only tackle discretized values. Traced time segments problem has already been a serious problem of data mining from a time series with rough set, so an innovative method to determine the traced time segments is proposed. Finally, two mining strategies are proposed to demonstrate the process of mining association rules in a time series with rough set, and an example is presented too