Time series Analysis

For forecasting and time-series analysis PVStatLab applies advanced statistical techniques including the ARMA and SSA methodologies. One of examples is forecasting of power production from PV systems.



ARMA model and its generalizations

Autoregressive–moving-average (ARMA) models are a well-known tool for understanding and predicting future values in a series. ARMA models can be generalized in many ways such as autoregressive conditional heteroskedasticity (ARCH) models and autoregressive integrated moving average (ARIMA) models. If multiple time series are to be fitted then a vector ARIMA (or VARIMA) model may be fitted. If the data contains seasonal effects, it may be modeled by a SARIMA (seasonal ARIMA) or a periodic ARMA model.


Singular Spectrum Analysis

Singular Spectrum Analysis (SSA) is a sophisticated method of time series analysis and forecasting. It combines advantages of many classical methods, such as Fourier and regression analyses. The basic SSA algorithm for analyzing time series consists of:

  • Transformation of the time series into a trajectory matrix using the moving window; 
  • Singular Value Decomposition (SVD) of this matrix; 
  • Reconstruction of a time series based on selected eigentriples.