Package contains methods for time series representations computation. Representation methods of time series are for dimensionality and noise reduction, emphasizing of main characteristics of time series data and speed up of consequent usage of machine learning methods.
| Package: | TSrepr |
| Type: | Package |
| Date: | 2018-01-26 - Inf |
| License: | GPL-3 |
The following functions for time series representations are included in the package:
repr_paa - Piecewise Aggregate Approximation (PAA)
repr_dwt - Discrete Wavelet Transform (DWT)
repr_dft - Discrete Fourier Transform (DFT)
repr_dct - Discrete Cosine Transform (DCT)
repr_sma - Simple Moving Average (SMA)
repr_pip - Perceptually Important Points (PIP)
repr_sax - Symbolic Aggregate Approximation (SAX)
repr_pla - Piecewise Linear Approximation (PLA)
repr_seas_profile - Mean seasonal profile
repr_lm - Model-based seasonal representations based on linear model (lm, rlm, l1)
repr_gam - Model-based seasonal representations based on generalized additive model (GAM)
repr_exp - Exponential smoothing seasonal coefficients
repr_feaclip - Feature extraction from clipping representation (FeaClip)
repr_featrend - Feature extraction from trending representation (FeaTrend)
repr_feacliptrend - Feature extraction from clipping and trending representation (FeaClipTrend)
There are also implemented additional useful functions as:
repr_windowing - applies above mentioned representations to every window of a time series
repr_matrix - applies above mentioned representations to every row of a matrix of time series
norm_z, norm_min_max - normalisation functions
norm_z_params, norm_min_max_params - normalisation functions with defined parameters
norm_z_list, norm_min_max_list - normalisation functions with output also of scaling parameters
denorm_z, denorm_min_max - denormalisation functions