spkit.TD_Embed

spkit.TD_Embed(x, order=3, delay=1)

Time delay Embedding Matrix

Extracting Embeddings

Parameters:

x1d-array

• time series of shape (n,)

orderint, default=3

• Embedding dimension (order).

delayint, default=1

• Delay.

Returns:

X: Embedded Matrix: ndarray

• Embedded time-series, of shape (n - (order - 1) * delay, order)

See also

signal_embeddings

Signal Embeddings

create_multidim_space_signal

Creating Multi-Dimensional Signal Space

Examples

#sp.TD_Embed
import numpy as np
import matplotlib.pyplot as plt
import spkit as sp
X, fs, ch_names = sp.data.eeg_sample_14ch()
x1 = X[500:700,0]
t = np.arange(len(x1))/fs
Xe = sp.TD_Embed(x1,order=3,delay=2)

plt.figure(figsize=(10,4))
plt.subplot(211)
plt.plot(t,x1)
plt.xlim([t[0],t[-1]])
plt.ylabel('x')
plt.xticks([])
plt.subplot(212)
plt.plot(t[:Xe.shape[0]],Xe)
plt.xlim([t[0],t[-1]])
plt.ylabel('Embeddings')
plt.xlabel('time (s)')
idx = 47
plt.axvline(t[idx],color='k',lw=1,alpha=0.5)
plt.plot([t[idx],t[idx],t[idx]],Xe[idx],'o',ms=3)
plt.plot(t[[idx,idx+5,idx+10]],Xe[idx]+50,ls='-',marker='.',color='k')
plt.text(t[idx+10],80,' embedding')
plt.tight_layout()
plt.show()