-- numpy, pylab
The pylab output will draw
-- a set of points inside a circle defined by x0,y0,R0
-- the circle (x0,y0) with rad R0
-- the optimized circle with minimum R enclosing the points
"""
from mystic.models import circle, sparse_circle
import pylab
# generate training set & define cost function
# CostFactory2 allows costfunction to reuse datapoints from training set
x0, y0, R0 = [10., 20., 3]
npts = 20
xy = sparse_circle(x0, y0, R0, npts)
cost = sparse_circle.CostFactory2(xy)
# function to find the 'support vectors' given R
# SV in quotes because they are found by optimizing the primal,
# as opposed to "directly" via the dual.
def sv(data, xx,yy,rr):
svl = []
for i in range(len(data)):
x,y = data[i]
if abs((xx-x)*(xx-x)+(yy-y)*(yy-y) - rr*rr) < 0.01:
svl.append(i)
return svl
# DEsolver inputs
MAX_GENERATIONS = 2000
-- numpy, pylab
The pylab output will draw
-- a set of points inside a circle defined by x0,y0,R0
-- the circle (x0,y0) with rad R0
-- the optimized circle with minimum R enclosing the points
"""
from mystic.models import circle, sparse_circle
import pylab
# generate training set & define cost function
# CostFactory2 allows costfunction to reuse datapoints from training set
x0, y0, R0 = [10., 20., 3]
npts = 20
xy = sparse_circle(x0, y0, R0, npts)
cost = sparse_circle.CostFactory2(xy)
# function to find the 'support vectors' given R
# SV in quotes because they are found by optimizing the primal,
# as opposed to "directly" via the dual.
def sv(data, xx,yy,rr):
svl = []
for i in range(len(data)):
x,y = data[i]
if abs((xx-x)*(xx-x)+(yy-y)*(yy-y) - rr*rr) < 0.01:
svl.append(i)
return svl
# DEsolver inputs
MAX_GENERATIONS = 2000