In [1]:

    

from __future__ import division
import jedi.jedi as jedi
from jedi.utils import plot, init_tools
import cPickle
import seaborn as sns
import matplotlib.pylab as plt
import numpy as np
%config InlineBackend.figure_format = 'retina'
%matplotlib inline


    

    




/Users/simonhaxby/anaconda/envs/py27/lib/python2.7/site-packages/matplotlib/__init__.py:872: UserWarning: axes.color_cycle is deprecated and replaced with axes.prop_cycle; please use the latter.
  warnings.warn(self.msg_depr % (key, alt_key))

In [2]:

    

noiseless = cPickle.load(open("../data/randompickles/noiseless_errors.p", "rb"))
noise = cPickle.load(open("../data/randompickles/noise_errors.p", "rb"))


    

In [3]:

    

noise.keys()


    

    
Out[3]:





[0.25, 0.2, 0.05, 0.15, 0.3, 0.1]

In [4]:

    

plt.figure(figsize=(8,10))
plt.subplot(2,1,1)
for errors in noiseless['force'][0]:
    sns.distplot(errors, kde=False)
plt.subplot(2,1,2)
for errors in noiseless['dforce'][0]:
    sns.distplot(errors, kde=False)


    

    




/Users/simonhaxby/anaconda/envs/py27/lib/python2.7/site-packages/numpy/lib/function_base.py:564: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future
  n = np.zeros(bins, ntype)
/Users/simonhaxby/anaconda/envs/py27/lib/python2.7/site-packages/numpy/lib/function_base.py:611: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future
  n += np.bincount(indices, weights=tmp_w, minlength=bins).astype(ntype)






    

In [103]:

    

print "Noiseless --"
print"FORCE: %.4f"% (np.var(noiseless['force'][0]))
print"DFORCE: %.4f"% (np.var(noiseless['dforce'][0]))


    

    




Noiseless --
FORCE: 0.2998
DFORCE: 0.4574

In [12]:

    

np.concatenate(noiseless['dforce'][0])


    

    
Out[12]:





(15030,)

In [97]:

    

level = .2


    

In [98]:

    

plt.figure(figsize=(8,10))
plt.subplot(2,1,1)
for errors in noise[level]['force'][0]:
    sns.distplot(errors, kde=False)
plt.subplot(2,1,2)
for errors in noise[level]['dforce'][0]:
    sns.distplot(errors, kde=False)


    

In [5]:

    

dns = []
fns = []

print "Gaussian Noise --\n"
for level in sorted(noise.keys()):
    print "Variance: %.2f" % level
    print"FORCE: %.4f"% (np.var(noise[level]['force'][0]))
    print"DFORCE: %.4f\n"% (np.var(noise[level]['dforce'][0]))
    dns.append(np.var(noise[level]['dforce'][0]))
    fns.append(np.var(noise[level]['force'][0]))


    

    




Gaussian Noise --

Variance: 0.05
FORCE: 0.1648
DFORCE: 0.1029

Variance: 0.10
FORCE: 0.2019
DFORCE: 0.1289

Variance: 0.15
FORCE: 0.2832
DFORCE: 0.2247

Variance: 0.20
FORCE: 0.4137
DFORCE: 0.3934

Variance: 0.25
FORCE: 0.4146
DFORCE: 0.3355

Variance: 0.30
FORCE: 0.4698
DFORCE: 0.4615

In [16]:

    

sns.set_style('white')
plt.figure(figsize=(8,5))
plt.plot(noise.keys(), dns, 'go', label='DFORCE');
plt.plot(noise.keys(), fns, 'bo', label='FORCE');
plt.xlim(0,.35);
plt.ylim(0,1)
plt.xlabel("$\sigma$", fontsize=30);
plt.ylabel("MSE", fontsize=25);
plt.legend(prop={'size':20});


    

In [ ]: