A new pseudo-parametric approach for unit root test using singular spectrum analysis and Gini-Ridge regression
نویسندگان
1 The Research Institute of Energy Management and Planning, University of Tehran, Tehran, Ira
2 Department of Statistics, University of Payame Noor, Tehran, Iran
3 Department of Statistics, University of Payame Noor, Tehran, Iran
4 Department of Statistics, University of Payame Noor, Tehran, Iran
doi
10.22034/jsmta.2025.21935.1149چکیده
This article presents a novel method for conducting a unit root test utilizing a combination of singular spectrum analysis and the Gini-Ridge approach. The Gini-Ridge approach employs a pseudo-parametric Gini-Ridge regression technique to estimate the coefficient from the AR(1) model while also utilizing bootstrap resampling to determine critical values for the test. The effectiveness of the Gini Ridge-based stationarity test is evaluated across various probability distributions, including the normal, Laplace, skew-normal, and Cauchy distributions. Simulation results validate the proposed test and demonstrate its superiority of over existing procedures. In addition, the proposed method for forecasting the number mobile transactions conducted by all e-commerce companies operating in the industry, under the supervision of Shaparak and the Central Bank of Iran, from April 2019 to August 2024, was evaluated. The findings showed that this new approach has a superior ability to detect the presence of unit roots in time series data.
