if model.lower() == 'kf':
p = 10
kf = KalmanFilter(data_dict, p)
kf.fit()
kf.display(out="./mg_kf.png")
if model.lower() == "ar":
p = 50
my_ar = AutoRegressive(data_dict, p)
my_ar.fit()
my_ar.predict()
my_ar.display(out="./mg_ar.png")
if model.lower() == "ac":
p = 50
my_ac = AutoCorrelation(data_dict, p)
my_ac.fit()
my_ac.predict()
my_ac.display(out="./mg_ac.png")
my_ac.spectrum()
my_ac.plot_attr("spectrum", show=True)
if model.lower() == "gp":
Q = 3
use_kernels = "exponential_quadratic* cosine"
for _ in range(Q - 1):
use_kernels += "+ exponential_quadratic * cosine"
# use_kernels = 'rational_quadratic + periodic'
use_means = "constant"
if model.lower() == 'kf':
p = 25
kf = KalmanFilter(data_dict, p)
kf.fit()
kf.display(out="./co2_kf.png")
if model.lower() == "ar":
p = 50
my_ar = AutoRegressive(data_dict, p)
my_ar.fit()
my_ar.predict()
my_ar.display(out="./co2_ar.png")
if model.lower() == "ac":
p = 50
my_ac = AutoCorrelation(data_dict, p)
my_ac.fit()
my_ac.predict()
my_ac.display(out="./co2_ac.png")
my_ac.spectrum()
if model.lower() == "gp":
Q = 3
use_kernels = "exponential_quadratic* cosine"
for _ in range(Q - 1):
use_kernels += "+ exponential_quadratic * cosine"
# use_kernels = 'rational_quadratic + periodic'
use_means = "constant"
estimator = "MLE"
from process_data import data_from_file
from Regression import AutoRegressive, AutoCorrelation
# file_name = "finPredProb.mat"
# file_name = "co2.mat"
# file_name = "sunspots.mat"
# file_name = "mg.mat"
file_name = "fXSamples.mat"
ix = 1
p = 5
args = data_from_file(file_name,
ix=ix)
my_ar = AutoRegressive(*args, p=p)
my_ar.fit()
my_ar.predict()
# my_ar.plot_var('ypred')
my_ac = AutoCorrelation(*args, p=p)
my_ac.fit()
my_ac.predict()
# my_ac.plot_var('ypred', show=True)
my_ac.spectrum()
my_ac.plot_attr('spectrum', show=True)
