Time Series To Features¶
Converts time-series of length N into a target signal ‘f0’ and M features signals ‘f1 .. fM’ of length (N-M)
Documentation¶
Converts time-series of length N into a target signal ‘f0’ and M features signals ‘f1 .. fM’ of length (N-M) as follows:
f0 = x_{M+1} .. x_{n}
f1 = x_1, … x_{n-m}
f2 = x_2, … x_{n-m+1}
…
fM = x_M, … x_{n-1}
The features ‘f1 .. fM’ thus capture the last M values of the signal f0.
Definition¶
Input ports¶
- input table
Input signals
Output ports¶
- x table
Features
- y table
Target signal
Configuration¶
- History length (M)
Number of points of history (M) captured at each time point
Examples¶
Implementation¶
- class node_difference.TimeSeriesToFeatures[source]