Discretized function representation

Shows how to make a discretized representation of a function.

# Author: Carlos Ramos Carreño <vnmabus@gmail.com>
# License: MIT

# sphinx_gallery_thumbnail_number = 2

We will construct a dataset containing several sinusoidal functions with random displacements.

import numpy as np

random_state = np.random.RandomState(0)

grid_points = np.linspace(0, 1)
data = np.array([
    np.sin((grid_points + random_state.randn()) * 2 * np.pi)
    for _ in range(5)
])

The FDataGrid class is used for datasets containing discretized functions that are measured at the same points.

from skfda import FDataGrid

fd = FDataGrid(
    data,
    grid_points,
    dataset_name="Sinusoidal curves",
    argument_names=["t"],
    coordinate_names=["x(t)"],
)

fd = fd[:5]

We can plot the measured values of each function in a scatter plot.

import matplotlib.pyplot as plt

fd.scatter(s=0.5)
plt.show()

We can also plot the interpolated functions.

fd.plot()
plt.show()