حسین حق بین

In this dissertation, the first general algorithm of singular spectrum analysis (SSA) is introduced as a nonparametric method in time series analysis, and then the generalizations of a multivariable SSA method are presented (MSSA). The generalization of this method to the two-dimensional mode (2D-SSA) and the mode that can be used to analyze images with any arbitrary shape (ShSSA) is also investigated. It is also shown that the ShSSA method is a generalization which includes any other generalization of the SSA method as a specific case. In each of the above generalizations, it attempt to apply the mentioned methods in real data and simulation studies to demonstrate the efficiency of the methods.