notebook.community

Edit and run



In [7]:

    
from data.complex.PCAD_main import nonLinearPCA, linearPCA
from data.complex.SolverNL import SolverDiff, lambdify, solver
from data.complex.complex import (
    examples, 
    polynomial, 
    plotSymbolicEq, 
    plotInComplexNumbers, 
    makeToComplexForm
    ) 

from data.finance.addins.neuralNetTestModified import NNtrainer, NeuralNetForward
from data.finance.addins.mlToFinance1 import LinkQuandl
from data.finance.addins.main import (
    plotSleepStudy, 
    generateSampleData, 
    testNeuralNetForward, 
    testLogic )

import matplotlib.pyplot as plt
import numpy as np



In [8]:

    
lq = LinkQuandl()
testLogic(lq)

# Neural network    
x = NeuralNetForward.x/np.amax(NeuralNetForward.x, axis = 0)
y = NeuralNetForward.y/100
testX = NeuralNetForward.testX/np.amax(NeuralNetForward.testX, axis = 0)
testY = NeuralNetForward.testY/100   
plotSleepStudy(x, y)
plotSleepStudy(testX, testY)
    
allInputs, hoursSleep, hoursStudy = generateSampleData (sampleSize = 100, max = 10, min = 0)
nn = NeuralNetForward(x = x, y = y)
#testNeuralNetForward(nn, x, y)









    



x array:  [  0.   1.   2.   3.   4.   5.   6.   7.   8.   9.  10.  11.  12.  13.  14.
  15.  16.  17.  18.  19.  20.  21.  22.  23.  24.  25.  26.  27.  28.  29.
  30.  31.  32.  33.  34.  35.  36.  37.  38.  39.] 
y array [  60.  -43.   63.   41.   14.  -39.   -2.  -38.  -16.   35.   65.   21.
   62.   41.  -13.   44.   22.   52.   33.    9.   69.   40.   83.    1.
   16.    9.   -6.   25.  116.  103.  108.  121.   22.   48.   19.  106.
   55.  132.  119.   90.]
predict y:  21.357879925
coefficientDet:  0.308484102555
Coefficience:  0.308484102555






    












    












    












    



NeuralNetForward initialized with 
input [[ 0.3  1. ]
 [ 0.5  0.2]
 [ 1.   0.4]
 [ 0.7  0.6]] 
output[[ 0.75]
 [ 0.82]
 [ 0.93]
 [ 0.87]]



In [5]:

    
t, v0, g, dydt = SolverDiff.symbolicMathTest()
v = lambdify([t, v0, g], dydt)
    
roots = SolverDiff.solverUsingDiff()
f = SolverDiff.taylorSeries()
    
roots









    



y:  -g*t**2/2 + t*v0
derivative of: dy/dt:  -g*t + v0
second derivative:  -g
y integral:  -g*t**3/6 + t**2*v0/2
0
1 + t + \frac{t^{2}}{2} + \mathcal{O}\left(t^{3}\right)
1 + t + \frac{t^{2}}{2} - \frac{t^{4}}{8} - \frac{t^{5}}{15} - \frac{t^{6}}{240} + \frac{t^{7}}{90} + \mathcal{O}\left(t^{8}\right)






    Out[5]:





[0, 2*v0/g]



In [9]:

    
solver()









    



0:  |F(x)| = 626.642; step 1; tol 0.000157679
1:  |F(x)| = 234.793; step 1; tol 0.126349
2:  |F(x)| = 114.959; step 1; tol 0.215755
3:  |F(x)| = 24.1208; step 1; tol 0.0396221
4:  |F(x)| = 1.53221; step 1; tol 0.00363161
5:  |F(x)| = 0.118354; step 1; tol 0.00536992
6:  |F(x)| = 0.00888137; step 1; tol 0.00506802
7:  |F(x)| = 0.000125105; step 1; tol 0.000178578
Residual: 5.80521e-06



In [10]:

    
linearPCA()









    












    



[ 0.75871884  0.01838551]
[[ 0.94446029  0.32862557]
 [ 0.32862557 -0.94446029]]






    












    



(200, 2)
(200, 1)



In [11]:

    
nonLinearPCA()



In [26]:

    
examples()
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1)
circ = plt.Circle((0.5, 0.5), radius=0.5, edgecolor='b', facecolor='None')
ax.add_patch(circ)
plt.show()
polynomial()
plotSymbolicEq()
    
a = np.arange(5) + 1j*np.arange(6,11)
plotInComplexNumbers(a)
    
f = np.logspace(-2,4,10)
y = makeToComplexForm(f)









    



real part:  2.0
imag part:  3.0
Squar:  (-5+12j)
Distance:  3.605551275463989
cmath sin:  (9.15449914691143-4.168906959966565j)






    












    



a*x**2 + b*x + c
\frac{1}{2 a} \left(- b + \left(- 4 a c + b^{2}\right)^{0.5}\right)






    












    












    



[  1.00000000e-02   4.64158883e-02   2.15443469e-01   1.00000000e+00
   4.64158883e+00   2.15443469e+01   1.00000000e+02   4.64158883e+02
   2.15443469e+03   1.00000000e+04]
[ 1. +6.28318531e-02j  1. +2.91639628e-01j  1. +1.35367124e+00j
  1. +6.28318531e+00j  1. +2.91639628e+01j  1. +1.35367124e+02j
  1. +6.28318531e+02j  1. +2.91639628e+03j  1. +1.35367124e+04j
  1. +6.28318531e+04j]



In [ ]: