Singular spectrum analysis and forecasting of failure time series

Singular spectrum analysis (SSA) is a relatively recent approach used to model time series with no assumptions of the underlying process. SSA is able to make a decomposition of the original time series into the sum of independent components, which represent the trend, oscillatory behavior (periodic or quasi-periodic components) and noise. In this paper SSA is used to decompose and forecast failure behaviors using time series related to time-to-failure data. Results are compared with previous approaches and show that SSA is a promising approach for data analysis and for forecasting failure time series.