temporal_PCA: Temporal PCA

In canhakan/canhakan1: Helper Package for Spatio-temporal Time-Series Data

Description

A function that transforms a data with lag values to a format where multiplication with eigenvectors of temporal covariance is possible. Then multiplicates with the eigenvectors matrix and returns the dimension-reduced data.

Usage

temporal_PCA(data, lags, features, ev, n)

Arguments

data	data with lag values. Columnnames must be of format (featurename)_(lag) (e.g. ne_lag3, nw_lead1, se)
lags	list of lags to be used. Same as create_lagged() argument
features	list of feature names without lag/lead additions
ev	the temporal eigenvector matrix
n	represents how many Principal Components to be returned

Value

Time dimension reduced data. (e.g. ne_Time1, sw_Time2)

canhakan/canhakan1 documentation built on Dec. 19, 2021, 1:48 p.m.