linearity_demonstration

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Wed Nov 13 17:52:03 2019

@author: thausmann
"""
import numpy
import matplotlib.pyplot as plt
import example
import ML_service_recomendor

def pyt(x, y):
    return (x**2 + y**2)**(1/2)

def unit_circle(x, y):
    length = pyt(x, y)
    return x/length, y/length

Ax, Ay = [numpy.random.normal()/6 for i in range(100)], [numpy.random.normal()/6 for i in range(100)]

Bx, By = [], []
for x, y in zip(Ax, Ay):
    nx, ny = [e-q for e, q in zip(unit_circle(x,y),(x,y))]
    Bx.append(nx)
    By.append(ny)
    

PointsA = (Ax,Ay,'ro')
PointsB = (Bx,By,'bo')

generator = ML_service_recomendor.train([[Axi,Ayi] for Axi,Ayi in zip(Ax,Ay)]
                                    +[[Bxi,Byi] for Bxi,Byi in zip(Bx,By)],
                                    [[0,1] if i<250 else [1,0] for i in range(499)])

Models = []
for Model in generator:
    Models.append(Model)

i = max(int(len(Models)/5),1)

for Model in Models[::i]:
    example.GRAPH(lambda x,y: Model([x**2,y**2,x,y]), PointsA, PointsB)