def test_pvalues_of_F():
#According to a great online F distribution calculator:
# If X = F(3,12), P(X<1.75) = .79
# P(X< 0.5) = 0.31
sigFig = 2
F = stats.f( dimension, sample_size-dimension)
print almost_equal(f.cdf(1.75), 0.79, sigFig)
print almost_equal(f.cdf(0.5), 0.31, sigFig)
def inv_safe(data):
'''I was having some strange things happening where matrices that should be
invertible were having incorrect inverses.
'''
eps =1e-2
d_sinv = sp.linalg.inv(data)
#print data*d_sinv
return almost_equal(sp.dot(data,d_sinv), np.eye(data.shape[0]))
def test_inv():
print "Test that inverse is working:"
x = True
#a,b,c,d,e,f,g,h,i=range(1,10)
#a*e*i - a*f*h - b*d*i + b*f*g + c*d*h - c*e*g
for i in range(5):
data = np.random.rand(3,3) * 10
#print "next"
#print np.dot(sp.linalg.invhilbert(data), data)
#print np.dot(sp.linalg.inv(data), data)
#print sp.linalg.inv(data) * data
#print np.dot(d_inv, data)
if not inv_safe(data) and not almost_equal(sp.linalg.det(data),0.0) :
print "broken"
x = False
if x:
print "Inverse is working"
