According to (steward,1998). A matrix A which is invertible can be approximated by the formula A^{-1} = \sum^{inf}_{n=0} (I- A)^{n}
I tried implementing an algorithm to approximate a simple matrix's inverse, the loss function showed funny results. please look at the code below. more info about the Neumann series can be found here and here
here is my code.
A = np.array([[1,0,2],[3,1,-2],[-5,-1,9]])
class Neumann_inversion():
def __init__(self,A,rank):
self.A = A
self.rank = rank
self.eye = np.eye(len(A))
self.loss = []
self.loss2 =[]
self.A_hat = np.zeros((3,3),dtype = float)
#self.loss.append(np.linalg.norm(np.linalg.inv(self.A)-self.A_hat))
def approximate(self):
# self.A_hat = None
n = 0
L = (self.eye-self.A)
while n < self.rank:
self.A_hat += np.linalg.matrix_power(L,n)
loss = np.linalg.norm(np.linalg.inv(self.A) - self.A_hat)
self.loss.append(loss)
n+= 1
plt.plot(self.loss)
plt.ylabel('Loss')
plt.xlabel('rank')
# ax.axis('scaled')
return
Matrix = Neumann_inversion(A,200)
Matrix.approximate()
The formula is valid only if $A^n$ tends to zero as $n$ increase. So your matrix must satisfy
np.all(np.abs(np.linalg.eigvals(A)) < 1)
Try
Neumann_inversion(A/10, 200).approximate()
and you can take the loss seriously :)
The origin of the formula has something to do with
(1-x) * (1 + x + x^2 + ... x^n) = (1 - x^(n+1))
If, and only if, all the eigenvalues of the matrix have magnitude less than 1 the term x^(n+1) will be close to zero, so the sum will be approximately the inverse of (1-x).
I do have a function, for example , but this can be something else as well, like a quadratic or logarithmic function. I am only interested in the domain of . The parameters of the function (a and k in this case) are known as well.
My goal is to fit a continuous piece-wise function to this, which contains alternating segments of linear functions (i.e. sloped straight segments, each with intercept of 0) and constants (i.e. horizontal segments joining the sloped segments together). The first and last segments are both sloped. And the number of segments should be pre-selected between around 9-29 (that is 5-15 linear steps + 4-14 constant plateaus).
Formally
The input function:
The fitted piecewise function:
I am looking for the optimal resulting parameters (c,r,b) (in terms of least squares) if the segment numbers (n) are specified beforehand.
The resulting constants (c) and the breakpoints (r) should be whole natural numbers, and the slopes (b) round two decimal point values.
I have tried to do the fitting numerically using the pwlf package using a segmented constant models, and further processed the resulting constant model with some graphical intuition to "slice" the constant steps with the slopes. It works to some extent, but I am sure this is suboptimal from both fitting perspective and computational efficiency. It takes multiple minutes to generate a fitting with 8 slopes on the range of 1-50000. I am sure there must be a better way to do this.
My idea would be to instead using only numerical methods/ML, the fact that we have the algebraic form of the input function could be exploited in some way to at least to use algebraic transforms (integrals) to get to a simpler optimization problem.
import numpy as np
import matplotlib.pyplot as plt
import pwlf
# The input function
def input_func(x,k,a):
return np.power(x,1/a)*k
x = np.arange(1,5e4)
y = input_func(x, 1.8, 1.3)
plt.plot(x,y);
def pw_fit(func, x_r, no_seg, *fparams):
# working on the specified range
x = np.arange(1,x_r)
y_input = func(x, *fparams)
my_pwlf = pwlf.PiecewiseLinFit(x, y_input, degree=0)
res = my_pwlf.fit(no_seg)
yHat = my_pwlf.predict(x)
# Function values at the breakpoints
y_isec = func(res, *fparams)
# Slope values at the breakpoints
slopes = np.round(y_isec / res, decimals=2)
slopes = slopes[1:]
# For the first slope value, I use the intersection of the first constant plateau and the input function
slopes = np.insert(slopes,0,np.round(y_input[np.argwhere(np.diff(np.sign(y_input - yHat))).flatten()[0]] / np.argwhere(np.diff(np.sign(y_input - yHat))).flatten()[0], decimals=2))
plateaus = np.unique(np.round(yHat))
# If due to rounding slope values (to two decimals), there is no change in a subsequent step, I just remove those segments
to_del = np.argwhere(np.diff(slopes) == 0).flatten()
slopes = np.delete(slopes,to_del + 1)
plateaus = np.delete(plateaus,to_del)
breakpoints = [np.ceil(plateaus[0]/slopes[0])]
for idx, j in enumerate(slopes[1:-1]):
breakpoints.append(np.floor(plateaus[idx]/j))
breakpoints.append(np.ceil(plateaus[idx+1]/j))
breakpoints.append(np.floor(plateaus[-1]/slopes[-1]))
return slopes, plateaus, breakpoints
slo, plat, breaks = pw_fit(input_func, 50000, 8, 1.8, 1.3)
# The piecewise function itself
def pw_calc(x, slopes, plateaus, breaks):
x = x.astype('float')
cond_list = [x < breaks[0]]
for idx, j in enumerate(breaks[:-1]):
cond_list.append((j <= x) & (x < breaks[idx+1]))
cond_list.append(breaks[-1] <= x)
func_list = [lambda x: x * slopes[0]]
for idx, j in enumerate(slopes[1:]):
func_list.append(plateaus[idx])
func_list.append(lambda x, j=j: x * j)
return np.piecewise(x, cond_list, func_list)
y_output = pw_calc(x, slo, plat, breaks)
plt.plot(x,y,y_output);
(Not important, but I think the fitted piecewise function is not continuous as it is. Intervals should be x<=r1; r1<x<=r2; ....)
As Anatolyg has pointed out, it looks to me that in the optimal solution (for the function posted at least, and probably for any where the derivative is different from zero), the horizantal segments will collapse to a point or the minimum segment length (in this case 1).
EDIT---------------------------------------------
The behavior above could only be valid if the slopes could have an intercept. If the intercepts are zero, as posted in the question, one consideration must be taken into account: Is the initial parabolic function defined in zero or nearby? Imagine the function y=0.001 *sqrt(x-1000), then the segments defined as b*x will have a slope close to zero and will be so similar to the constant segments that the best fit will be just the line that without intercept that fits better all the function.
Provided that the function is defined in zero or nearby, you can start by approximating the curve just by linear segments (with intercepts):
divide the function domain in N intervals(equal intervals or whose size is a function of the average curvature (or second derivative) of the function along the domain).
linear fit/regression in each intervals
for each interval, if a point (or bunch of points) in the extreme of any interval is better fitted by the line of the neighbor interval than the line of its interval, this point is assigned to the neighbor interval.
Repeat from 2) until no extreme points are moved.
Linear regressions might be optimized not to calculate all the covariance matrixes from scratch on each iteration, but just adding the contributions of the moved points to the previous covariance matrixes.
Then each linear segment (LSi) is replaced by a combination of a small constant segment at the beginning (Cbi), a linear segment without intercept (Si), and another constant segment at the end (Cei). This segments are easy to calculate as Si will contain the middle point of LSi, and Cbi and Cei will have respectively the begin and end values of the segment LSi. Then the intervals of each segment has to be calculated as an intersection between lines.
With this, the constant end segment will be collinear with the constant begin segment from the next interval so they will merge, resulting in a series of constant and linear segments interleaved.
But this would be a floating point start solution. Next, you will have to apply all the roundings which will mess up quite a lot all the segments as the conditions integer intervals and linear segments without slope can be very confronting. In fact, b,c,r are not totally independent. If ci and ri+1 are known, then bi+1 is already fixed
If nothing is broken so far, the final task will be to minimize the error/cost function (I assume that it will be the integral of the error between the parabolic function and the segments). My guess is that gradients here will be quite a pain, as if you change for example one ci, all the rest of the bj and cj will have to adapt as well due to the integer intervals restriction. However, if you can generalize the derivatives between parameters ( how much do I have to adapt bi+1 if ci changes a unit), you can propagate the change of one parameter to all other parameters and have kind of a gradient. Then for each interval, you can estimate what would be the ideal parameter and averaging all intervals calculate the best gradient step. Let me illustrate this:
Assuming first that r parameters are fixed, if I change c1 by one unit, b2 changes by 0.1, c2 changes by -0.2 and b3 changes by 0.2. This would be the gradient.
Then I estimate, comparing with the parabolic curve, that c1 should increase 0.5 (to reduce the cost by 10 points), b2 should increase 0.2 (to reduce the cost by 5 points), c2 should increase 0.2 (to reduce the cost by 6 points) and b3 should increase 0.1 (to reduce the cost by 9 points).
Finally, the gradient step would be (0.5/1·10 + 0.2/0.1·5 - 0.2/(-0.2)·6 + 0.1/0.2·9)/(10 + 5 + 6 + 9)~= 0.45. Thus, c1 would increase 0.45 units, b2 would increase 0.45·0.1, and so on.
When you add the r parameters to the pot, as integer intervals do not have an proper derivative, calculation is not straightforward. However, you can consider r parameters as floating points, calculate and apply the gradient step and then apply the roundings.
We can integrate the squared error function for linear and constant pieces and let SciPy optimize it. Python 3:
import matplotlib.pyplot as plt
import numpy as np
import scipy.optimize
xl = 1
xh = 50000
a = 1.3
p = 1 / a
n = 8
def split_b_and_c(bc):
return bc[::2], bc[1::2]
def solve_for_r(b, c):
r = np.empty(2 * n)
r[0] = xl
r[1:-1:2] = c / b[:-1]
r[2::2] = c / b[1:]
r[-1] = xh
return r
def linear_residual_integral(b, x):
return (
(x ** (2 * p + 1)) / (2 * p + 1)
- 2 * b * x ** (p + 2) / (p + 2)
+ b ** 2 * x ** 3 / 3
)
def constant_residual_integral(c, x):
return x ** (2 * p + 1) / (2 * p + 1) - 2 * c * x ** (p + 1) / (p + 1) + c ** 2 * x
def squared_error(bc):
b, c = split_b_and_c(bc)
r = solve_for_r(b, c)
linear = np.sum(
linear_residual_integral(b, r[1::2]) - linear_residual_integral(b, r[::2])
)
constant = np.sum(
constant_residual_integral(c, r[2::2])
- constant_residual_integral(c, r[1:-1:2])
)
return linear + constant
def evaluate(x, b, c, r):
i = 0
while x > r[i + 1]:
i += 1
return b[i // 2] * x if i % 2 == 0 else c[i // 2]
def main():
bc0 = (xl + (xh - xl) * np.arange(1, 4 * n - 2, 2) / (4 * n - 2)) ** (
p - 1 + np.arange(2 * n - 1) % 2
)
bc = scipy.optimize.minimize(
squared_error, bc0, bounds=[(1e-06, None) for i in range(2 * n - 1)]
).x
b, c = split_b_and_c(bc)
r = solve_for_r(b, c)
X = np.linspace(xl, xh, 1000)
Y = [evaluate(x, b, c, r) for x in X]
plt.plot(X, X ** p)
plt.plot(X, Y)
plt.show()
if __name__ == "__main__":
main()
I have tried to come up with a new solution myself, based on the idea of #Amo Robb, where I have partitioned the domain, and curve fitted a dual - constant and linear - piece together (with the help of np.maximum). I have used the 1 / f(x)' as the function to designate the breakpoints, but I know this is arbitrary and does not provide a global optimum. Maybe there is some optimal function for these breakpoints. But this solution is OK for me, as it might be appropriate to have a better fit at the first segments, at the expense of the error for the later segments. (The task itself is actually a cost based retail margin calculation {supply price -> added margin}, as the retail POS software can only work with such piecewise margin function).
The answer from #David Eisenstat is correct optimal solution if the parameters are allowed to be floats. Unfortunately the POS software can not use floats. It is OK to round up c-s and r-s afterwards. But the b-s should be rounded to two decimals, as those are inputted as percents, and this constraint would ruin the optimal solution with long floats. I will try to further improve my solution with both Amo's and David's valuable input. Thank You for that!
import numpy as np
import matplotlib.pyplot as plt
from scipy.optimize import curve_fit
# The input function f(x)
def input_func(x,k,a):
return np.power(x,1/a) * k
# 1 / f(x)'
def one_per_der(x,k,a):
return a / (k * np.power(x, 1/a-1))
# 1 / f(x)' inverted
def one_per_der_inv(x,k,a):
return np.power(a / (x*k), a / (1-a))
def segment_fit(start,end,y,first_val):
b, _ = curve_fit(lambda x,b: np.maximum(first_val, b*x), np.arange(start,end), y[start-1:end-1])
b = float(np.round(b, decimals=2))
bp = np.round(first_val / b)
last_val = np.round(b * end)
return b, bp, last_val
def pw_fit(end_range, no_seg, **fparams):
y_bps = np.linspace(one_per_der(1, **fparams), one_per_der(end_range,**fparams) , no_seg+1)[1:]
x_bps = np.round(one_per_der_inv(y_bps, **fparams))
y = input_func(x, **fparams)
slopes = [np.round(float(curve_fit(lambda x,b: x * b, np.arange(1,x_bps[0]), y[:int(x_bps[0])-1])[0]), decimals = 2)]
plats = [np.round(x_bps[0] * slopes[0])]
bps = []
for i, xbp in enumerate(x_bps[1:]):
b, bp, last_val = segment_fit(int(x_bps[i]+1), int(xbp), y, plats[i])
slopes.append(b); bps.append(bp); plats.append(last_val)
breaks = sorted(list(x_bps) + bps)[:-1]
# If due to rounding slope values (to two decimals), there is no change in a subsequent step, I just remove those segments
to_del = np.argwhere(np.diff(slopes) == 0).flatten()
breaks_to_del = np.concatenate((to_del * 2, to_del * 2 + 1))
slopes = np.delete(slopes,to_del + 1)
plats = np.delete(plats[:-1],to_del)
breaks = np.delete(breaks,breaks_to_del)
return slopes, plats, breaks
def pw_calc(x, slopes, plateaus, breaks):
x = x.astype('float')
cond_list = [x < breaks[0]]
for idx, j in enumerate(breaks[:-1]):
cond_list.append((j <= x) & (x < breaks[idx+1]))
cond_list.append(breaks[-1] <= x)
func_list = [lambda x: x * slopes[0]]
for idx, j in enumerate(slopes[1:]):
func_list.append(plateaus[idx])
func_list.append(lambda x, j=j: x * j)
return np.piecewise(x, cond_list, func_list)
fparams = {'k':1.8, 'a':1.2}
end_range = 5e4
no_steps = 10
x = np.arange(1, end_range)
y = input_func(x, **fparams)
slopes, plats, breaks = pw_fit(end_range, no_steps, **fparams)
y_output = pw_calc(x, slopes, plats, breaks)
plt.plot(x,y_output,y);
I am trying to find the minimum of a natural cubic spline. I have written the following code to find the natural cubic spline. (I have been given test data and have confirmed this method is correct.) Now I can not figure out how to find the minimum of this function.
This is the data
xdata = np.linspace(0.25, 2, 8)
ydata = 10**(-12) * np.array([1,2,1,2,3,1,1,2])
This is the function
import scipy as sp
import numpy as np
import math
from numpy.linalg import inv
from scipy.optimize import fmin_slsqp
from scipy.optimize import minimize, rosen, rosen_der
def phi(x, xd,yd):
n = len(xd)
h = np.array(xd[1:n] - xd[0:n-1])
f = np.divide(yd[1:n] - yd[0:(n-1)],h)
q = [0]*(n-2)
for i in range(n-2):
q[i] = 3*(f[i+1] - f[i])
A = np.zeros(((n-2),(n-2)))
#define A for j=0
A[0,0] = 2*(h[0] + h[1])
A[0,1] = h[1]
#define A for j = n-2
A[-1,-2] = h[-2]
A[-1,-1] = 2*(h[-2] + h[-1])
#define A for in the middle
for j in range(1,(n-3)):
A[j,j-1] = h[j]
A[j,j] = 2*(h[j] + h[j+1])
A[j,j+1] = h[j+1]
Ainv = inv(A)
B = Ainv.dot(q)
b = (n)*[0]
b[1:(n-1)] = B
# now we find a, b, c and d
a = [0]*(n-1)
c = [0]*(n-1)
d = [0]*(n-1)
s = [0]*(n-1)
for r in range(n-1):
a[r] = 1/(3*h[r]) * (b[r + 1] - b[r])
c[r] = f[r] - h[r]*((2*b[r] + b[r+1])/3)
d[r] = yd[r]
#solution 1 start
for m in range(n-1):
if xd[m] <= x <= xd[m+1]:
s = a[m]*(x - xd[m])**3 + b[m]*(x-xd[m])**2 + c[m]*(x-xd[m]) + d[m]
return(s)
#solution 1 end
I want to find the minimum on the domain of my xdata, so a fmin didn't work as you can not define bounds there. I tried both fmin_slsqp and minimize. They are not compatible with the phi function I wrote so I rewrote phi(x, xd,yd) and added an extra variable such that phi is phi(x, xd,yd, m). M indicates in which subfunction of the spline we are calculating a solution (from x_m to x_m+1). In the code we replaced #solution 1 by the following
# solution 2 start
return(a[m]*(x - xd[m])**3 + b[m]*(x-xd[m])**2 + c[m]*(x-xd[m]) + d[m])
# solution 2 end
To find the minimum in a domain x_m to x_(m+1) we use the following code: (we use an instance where m=0, so x from 0.25 to 0.5. The initial guess is 0.3)
fmin_slsqp(phi, x0 = 0.3, bounds=([(0.25,0.5)]), args=(xdata, ydata, 0))
What I would then do (I know it's crude), is iterate this with a for loop to find the minimum on all subdomains and then take the overall minimum. However, the function fmin_slsqp constantly returns the initial guess as the minimum. So there is something wrong, which I do not know how to fix. If you could help me this would be greatly appreciated. Thanks for reading this far.
When I plot your function phi and the data you feed in, I see that its range is of the order of 1e-12. However, fmin_slsqp is unable to handle that level of precision and fails to find any change in your objective.
The solution I propose is scaling the return of your objective by the same order of precision like so:
return(s*1e12)
Then you get good results.
>>> sol = fmin_slsqp(phi, x0=0.3, bounds=([(0.25, 0.5)]), args=(xdata, ydata))
>>> print(sol)
Optimization terminated successfully. (Exit mode 0)
Current function value: 1.0
Iterations: 2
Function evaluations: 6
Gradient evaluations: 2
[ 0.25]
My implementation of steepest descent for solving Ax = b is showing some weird behavior: for any matrix large enough (~10 x 10, have only tested square matrices so far), the returned x contains all huge values (on the order of 1x10^10).
def steepestDescent(A, b, numIter=100, x=None):
"""Solves Ax = b using steepest descent method"""
warnings.filterwarnings(action="error",category=RuntimeWarning)
# Reshape b in case it has shape (nL,)
b = b.reshape(len(b), 1)
exes = []
res = []
# Make a guess for x if none is provided
if x==None:
x = np.zeros((len(A[0]), 1))
exes.append(x)
for i in range(numIter):
# Re-calculate r(i) using r(i) = b - Ax(i) every five iterations
# to prevent roundoff error. Also calculates initial direction
# of steepest descent.
if (numIter % 5)==0:
r = b - np.dot(A, x)
# Otherwise use r(i+1) = r(i) - step * Ar(i)
else:
r = r - step * np.dot(A, r)
res.append(r)
# Calculate step size. Catching the runtime warning allows the function
# to stop and return before all iterations are completed. This is
# necessary because once the solution x has been found, r = 0, so the
# calculation below divides by 0, turning step into "nan", which then
# goes on to overwrite the correct answer in x with "nan"s
try:
step = np.dot(r.T, r) / np.dot( np.dot(r.T, A), r )
except RuntimeWarning:
warnings.resetwarnings()
return x
# Update x
x = x + step * r
exes.append(x)
warnings.resetwarnings()
return x, exes, res
(exes and res are returned for debugging)
I assume the problem must be with calculating r or step (or some deeper issue) but I can't make out what it is.
The code seems correct. For example, the following test work for me (both linalg.solve and steepestDescent give the close answer, most of the time):
import numpy as np
n = 100
A = np.random.random(size=(n,n)) + 10 * np.eye(n)
print(np.linalg.eig(A)[0])
b = np.random.random(size=(n,1))
x, xs, r = steepestDescent(A,b, numIter=50)
print(x - np.linalg.solve(A,b))
The problem is in the math. This algorithm is guaranteed to converge to the correct solution if A is positive definite matrix. By adding the 10 * identity matrix to a random matrix, we increase the probability that all the eigen-values are positive
If you test with large random matrices (for example A = random.random(size=(n,n)), you are almost certain to have a negative eigenvalue, and the algorithm will not converge.
In R there is a very useful function that helps with determining parameters for a two sided t-test in order to obtain a target statistical power.
The function is called power.prop.test.
http://stat.ethz.ch/R-manual/R-patched/library/stats/html/power.prop.test.html
You can call it using:
power.prop.test(p1 = .50, p2 = .75, power = .90)
And it will tell you n the sample size needed to obtain this power. This is extremely useful in deterring sample sizes for tests.
Is there a similar function in the scipy package?
I've managed to replicate the function using the below formula for n and the inverse survival function norm.isf from scipy.stats
from scipy.stats import norm, zscore
def sample_power_probtest(p1, p2, power=0.8, sig=0.05):
z = norm.isf([sig/2]) #two-sided t test
zp = -1 * norm.isf([power])
d = (p1-p2)
s =2*((p1+p2) /2)*(1-((p1+p2) /2))
n = s * ((zp + z)**2) / (d**2)
return int(round(n[0]))
def sample_power_difftest(d, s, power=0.8, sig=0.05):
z = norm.isf([sig/2])
zp = -1 * norm.isf([power])
n = s * ((zp + z)**2) / (d**2)
return int(round(n[0]))
if __name__ == '__main__':
n = sample_power_probtest(0.1, 0.11, power=0.8, sig=0.05)
print n #14752
n = sample_power_difftest(0.1, 0.5, power=0.8, sig=0.05)
print n #392
Some of the basic power calculations are now available in statsmodels
http://statsmodels.sourceforge.net/devel/stats.html#power-and-sample-size-calculations
http://jpktd.blogspot.ca/2013/03/statistical-power-in-statsmodels.html
The blog article does not yet take the latest changes to the statsmodels code into account. Also, I haven't decided yet how many wrapper functions to provide, since many power calculations just reduce to the basic distribution.
>>> import statsmodels.stats.api as sms
>>> es = sms.proportion_effectsize(0.5, 0.75)
>>> sms.NormalIndPower().solve_power(es, power=0.9, alpha=0.05, ratio=1)
76.652940372066908
In R stats
> power.prop.test(p1 = .50, p2 = .75, power = .90)
Two-sample comparison of proportions power calculation
n = 76.7069301141077
p1 = 0.5
p2 = 0.75
sig.level = 0.05
power = 0.9
alternative = two.sided
NOTE: n is number in *each* group
using R's pwr package
> library(pwr)
> h<-ES.h(0.5,0.75)
> pwr.2p.test(h=h, power=0.9, sig.level=0.05)
Difference of proportion power calculation for binomial distribution (arcsine transformation)
h = 0.5235987755982985
n = 76.6529406106181
sig.level = 0.05
power = 0.9
alternative = two.sided
NOTE: same sample sizes
Matt's answer for getting the needed n (per group) is almost right, but there is a small error.
Given d (difference in means), s (standard deviation), sig (significance level, typically .05), and power (typically .80), the formula for calculating the number of observations per group is:
n= (2s^2 * ((z_(sig/2) + z_power)^2) / (d^2)
As you can see in his formula, he has
n = s * ((zp + z)**2) / (d**2)
the "s" part is wrong. a correct function that reproduces r's functionality is:
def sample_power_difftest(d, s, power=0.8, sig=0.05):
z = norm.isf([sig/2])
zp = -1 * norm.isf([power])
n = (2*(s**2)) * ((zp + z)**2) / (d**2)
return int(round(n[0]))
Hope this helps.
You also have:
from statsmodels.stats.power import tt_ind_solve_power
and put "None" in the value you want to obtain. For instande, to obtain the number of observations in the case of effect_size = 0.1, power = 0.8 and so on, you should put:
tt_ind_solve_power(effect_size=0.1, nobs1 = None, alpha=0.05, power=0.8, ratio=1, alternative='two-sided')
and obtain: 1570.7330663315456 as the number of observations required.
Or else, to obtain the power you can attain with the other values fixed:
tt_ind_solve_power(effect_size= 0.2, nobs1 = 200, alpha=0.05, power=None, ratio=1, alternative='two-sided')
and you obtain: 0.5140816347005553