Autocorrelation-based features

Computes various measures based on autocorrelation coefficients of the original series, first-differenced series and second-differenced series

Usage

acf_features(x)

Arguments

x

a univariate time series

Value

A vector of 6 values: first autocorrelation coefficient and sum of squared of first ten autocorrelation coefficients of original series, first-differenced series, and twice-differenced series. For seasonal data, the autocorrelation coefficient at the first seasonal lag is also returned.

Author

Thiyanga Talagala