MarketVector uses one particular analytical technique to first model and then project all time series published.

The analytical technique used is multi-channel singular spectrum analysis, or MSSA.  This analytical technique is a nonparametric spectral estimation method.

MSSA decomposes a time series into a trend component, many cyclical components, and noise.  Forecasting with MSSA entails finding a signal subspace that is, in turn, used to fit a linear recurrence relationship.  A forecast is essentially the continuation of the linear recurrence relationship.


MarketVector is produced entirely with the R programming language and many of the excellent packages that are available.  The multichannel singular spectrum analysis is performed using the excellent 'Rssa' package developed by Nina Golyandina and team.