Singular Spectrum Analysis and Its Application to the Industrial Inputs Price Index in Kenya

Abstract

The time-related features and future trajectory of any given time series may be better understood with the help of time series modelling and forecasting methods. The use of classical models without data transformation is hindered since the majority of financial and economic time series data does not adhere to the stringent assumptions of data normalcy, linearity, and stationarity. Because Singular Spectrum Analysis (SSA) is data-adaptive and non-parametric, it does not need to adhere to the same stringent assumptions as classical approaches. Using data collected from January 1992 through April 2022, this study examined the relationship between SSA and Kenya's monthly industrial inputs price index using a longitudinal research approach. In an effort to promote competitiveness and growth in Kenya's manufacturing sector, one of the country's manufacturing priorities since 2018 has been to lower the prices of industrial inputs. By 2022, the anticipated Manufacturing Value Added for Kenya was supposed to reach a whopping 22 percent. The research found that the SSA yields better accurate predictions (L = 12, r = 7, MAPE = 0.707%). According to the 24-period estimates, the cost of industrial inputs is expected to be reduced in the post-industrial agenda, although it is still high compared to 2017. So, we need to bring down the costs of industrial inputs to a reasonable level.