Clustering of unemployment rate data using a new independent component analysis algorithm

The performance of independent component analysis (ICA) algorithms is based on objective function and optimization algorithms and often their objective functions are based on dependency criteria. In this paper, to characterize the independence of two random variables, we present a dependence criteria dependency criterion based on copula and we study its estimated. Then, based on an estimator of the proposed dependency criterion, we present an algorithm for ICA, and the performance of the proposed algorithm is compared with some copula-based ICA algorithms, we used Amari error for comparison. Finally, we apply our method to one batch of real-time series for ICA as a pre-processing in clustering, and the obtained clusters are provided.

کلیدواژه ها:

Amari Error ، Clustering ، Copula ، Independent Components Analysis ، Unemployment Rate