example_narma.py

import pandas as pd
import numpy as np
import synthetica as sth


estimator = sth.NARMA


if __name__ == "__main__":

    
    # Model
    
    model = estimator(length=252, num_paths=1, seed=124)
    print(model)

    
    # Output
    
    df = model.transform()
    print(df)
    df.plot()
    # --matplotlib plt--

    
    # Underlying white noise
    
    noise = model.white_noise
    print(noise)
    # ...

    pd.DataFrame(noise).plot()
    # --matplotlib plt--

    
    # Testing callback
    
    # Mean
    # ----
    mean_value = model.mean
    # 0
    model.mean = 1
    noise = model.white_noise
    print(noise)
    # ...

    pd.DataFrame(noise).plot()
    # --matplotlib plt--

    # Delta
    # -----
    model.delta
    # 0.003968253968253968
    model.delta = 1/12
    noise = model.white_noise
    print(noise)
    # ...

    pd.DataFrame(noise).plot()
    # --matplotlib plt--

    # Sigma
    # -----
    model.sigma
    # 0.125
    model.sigma = 0.4
    noise = model.white_noise
    print(noise)
    # ...

    pd.DataFrame(noise).plot()
    # --matplotlib plt--

    
    # Cholesky
    

    model = estimator(num_paths=2, seed=123)
    # Create matrix for illustration purposes
    matrix = np.array([[1, .8], [.8, 1]])
    print(matrix)
    # ...
    
    
    # Without correlation
    
    df1 = model.transform()
    print(df1)
    # ...
    df1.plot()
    # --matplotlib plt--
    print(df1.corr())
    
    
    # With correlation
    
    df2 = model.transform(matrix)
    print(df2)
    # ...
    df2.plot()
    # --matplotlib plt--
    print(df2.corr())