feature visualization

Import related packages

import warnings
warnings.filterwarnings('ignore')
from scuteegfe import Feature
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import mne

generate siganls

data1 = np.empty((50, 5, 100))
for epoch in range(50):
    for ch in range(5):
        data1[epoch, ch] = np.linspace(ch, ch + epoch + 1, 100) + np.sin(np.linspace(-np.pi, np.pi, 100))

data2 = np.empty((50, 5, 100))
for epoch in range(50):
    for ch in range(5):
        data2[epoch, ch] = np.linspace(ch, ch + epoch + 1, 100) + np.random.randint(-100, 100) * 0.01 + np.sin(
            np.linspace(-np.pi, np.pi, 100)) + 5
        if ch == 3:
            data2[epoch, ch] = np.zeros(100)

all feature visualization

# calculate all feature
fea1 = Feature(data1, selected_funcs = Feature.funcs_subset_no_spect_slope).reorder().fix_missing()
fea2 = Feature(data2, selected_funcs = Feature.funcs_subset_no_spect_slope).reorder().fix_missing()
## boxplot
Feature.plot_feature_sns(fea1, fea2, ['ch1', 'ch2', 'ch3', 'ch4', 'ch5'])
plt.show()

## p value and heatmap
_, p = Feature.ttest_feature(fea1, fea2, ['ch1', 'ch2', 'ch3', 'ch4', 'ch5'])
plt.show()