Re: Determinant of Large Matrix
James Stroud j@mbia.edu wrote: James Stroud wrote: [pointless stuff] OK. Nevermind. I'm rebinding encodings and so taking a sample from the sample and thus getting the sample back. Terribly sorry. There is truly nothing to be sorry about. It takes guts to come right out and say that you made a mistake. Well done! start rant One of my pet peeves are people who pretend that they never make mistakes. Strangely enough, when you ask them to walk on water, they never quite manage it. This sort of thing should be cause for rejoicing, - when you think about it, its a proof that the body of knowledge that has to do with sampling is solid and reliable, independently of the experimenter - these days that is quite a comforting thought. The sanitised stuff you read in the scientific journals does not represent the true course of the progress of science - in reality, progress is the incidental by-product of a progressive series of blunders. But nobody involved will acknowledge this, for fear of their precious reputations... end rant - Hendrik -- http://mail.python.org/mailman/listinfo/python-list
Re: Determinant of Large Matrix
James Stroud wrote: If I run it again on 10 (or 1000) the set is basically homogenous but now of different values (terribly confusing): set([12048175104.1, 12048175104.15, 12048175104.46, 12048175103.94, 12048175104.23, 12048175103.81, 12048175103.98, 12048175103.9]) As you seem to have overread it: Note that Python only prints those numbers with full bogus precision since you let it display them using repr() (it's used in lists by default for display). If you use str() explicitly, Python applies rounding: A [12048175104.1, 12048175104.15, 12048175104.46, 12048175103.94, 12048175104.23, 12048175103.81, 12048175103.94, 12048175104.23, 12048175103.81, 12048175103.98, 12048175103.9] [repr(i) for i in A] ['12048175104.1', '12048175104.15', '12048175104.46', '12048175103.94', '12048175104.23', '12048175103.81', '12048175103.94', '12048175104.23', '12048175103.81', '12048175103.98', '12048175103.9'] [str(i) for i in A] ['12048175104.0', '12048175104.0', '12048175104.0', '12048175104.0', '12048175104.0', '12048175104.0', '12048175104.0', '12048175104.0', '12048175104.0', '12048175104.0', '12048175104.0'] Regards, Björn -- BOFH excuse #83: Support staff hung over, send aspirin and come back LATER. -- http://mail.python.org/mailman/listinfo/python-list
Re: Determinant of Large Matrix
Hi James Mathematica says that the determinant of the integer version of this matrix is 2774532096, which is another vote for the answer you have. Mathematica says that the determinant of the 24-digit real version of your matrix is 2.774532096*10^9, which looks very similar to me. I'd go with Numpy. Regards Mark Westwood PS 13x13 isn't a large matrix ! On Jun 6, 12:10 pm, James Stroud [EMAIL PROTECTED] wrote: Hello All, I'm using numpy to calculate determinants of matrices that look like this (13x13): [[ 0. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1.] [ 1. 0. 1. 4. 1. 9. 4. 4. 1. 1. 4. 9. 4. 9.] [ 1. 1. 0. 1. 4. 4. 9. 9. 4. 4. 1. 4. 1. 4.] [ 1. 4. 1. 0. 9. 1. 4. 4. 9. 1. 4. 1. 4. 1.] [ 1. 1. 4. 9. 0. 4. 4. 4. 1. 4. 1. 9. 4. 9.] [ 1. 9. 4. 1. 4. 0. 4. 4. 9. 4. 1. 1. 4. 1.] [ 1. 4. 9. 4. 4. 4. 0. 1. 1. 1. 9. 1. 9. 4.] [ 1. 4. 9. 4. 4. 4. 1. 0. 4. 1. 9. 4. 4. 1.] [ 1. 1. 4. 9. 1. 9. 1. 4. 0. 4. 4. 4. 4. 9.] [ 1. 1. 4. 1. 4. 4. 1. 1. 4. 0. 9. 4. 9. 4.] [ 1. 4. 1. 4. 1. 1. 9. 9. 4. 9. 0. 4. 1. 4.] [ 1. 9. 4. 1. 9. 1. 1. 4. 4. 4. 4. 0. 4. 1.] [ 1. 4. 1. 4. 4. 4. 9. 4. 4. 9. 1. 4. 0. 1.] [ 1. 9. 4. 1. 9. 1. 4. 1. 9. 4. 4. 1. 1. 0.]] For this matrix, I'm getting this with numpy: 2774532095.971 But I have a feeling I'm exceeding the capacity of floats here. Does anyone have an idea for how to treat this? Is it absurd to think I could get a determinant of this matrix? Is there a python package that could help me? Many thanks for any answers. James -- http://mail.python.org/mailman/listinfo/python-list
Determinant of Large Matrix
Hello All, I'm using numpy to calculate determinants of matrices that look like this (13x13): [[ 0. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1.] [ 1. 0. 1. 4. 1. 9. 4. 4. 1. 1. 4. 9. 4. 9.] [ 1. 1. 0. 1. 4. 4. 9. 9. 4. 4. 1. 4. 1. 4.] [ 1. 4. 1. 0. 9. 1. 4. 4. 9. 1. 4. 1. 4. 1.] [ 1. 1. 4. 9. 0. 4. 4. 4. 1. 4. 1. 9. 4. 9.] [ 1. 9. 4. 1. 4. 0. 4. 4. 9. 4. 1. 1. 4. 1.] [ 1. 4. 9. 4. 4. 4. 0. 1. 1. 1. 9. 1. 9. 4.] [ 1. 4. 9. 4. 4. 4. 1. 0. 4. 1. 9. 4. 4. 1.] [ 1. 1. 4. 9. 1. 9. 1. 4. 0. 4. 4. 4. 4. 9.] [ 1. 1. 4. 1. 4. 4. 1. 1. 4. 0. 9. 4. 9. 4.] [ 1. 4. 1. 4. 1. 1. 9. 9. 4. 9. 0. 4. 1. 4.] [ 1. 9. 4. 1. 9. 1. 1. 4. 4. 4. 4. 0. 4. 1.] [ 1. 4. 1. 4. 4. 4. 9. 4. 4. 9. 1. 4. 0. 1.] [ 1. 9. 4. 1. 9. 1. 4. 1. 9. 4. 4. 1. 1. 0.]] For this matrix, I'm getting this with numpy: 2774532095.971 But I have a feeling I'm exceeding the capacity of floats here. Does anyone have an idea for how to treat this? Is it absurd to think I could get a determinant of this matrix? Is there a python package that could help me? Many thanks for any answers. James -- http://mail.python.org/mailman/listinfo/python-list
Re: Determinant of Large Matrix
On 6 jun 2007, at 13.10, James Stroud wrote: Hello All, I'm using numpy to calculate determinants of matrices that look like this (13x13): [[ 0. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1.] [ 1. 0. 1. 4. 1. 9. 4. 4. 1. 1. 4. 9. 4. 9.] [ 1. 1. 0. 1. 4. 4. 9. 9. 4. 4. 1. 4. 1. 4.] [ 1. 4. 1. 0. 9. 1. 4. 4. 9. 1. 4. 1. 4. 1.] [ 1. 1. 4. 9. 0. 4. 4. 4. 1. 4. 1. 9. 4. 9.] [ 1. 9. 4. 1. 4. 0. 4. 4. 9. 4. 1. 1. 4. 1.] [ 1. 4. 9. 4. 4. 4. 0. 1. 1. 1. 9. 1. 9. 4.] [ 1. 4. 9. 4. 4. 4. 1. 0. 4. 1. 9. 4. 4. 1.] [ 1. 1. 4. 9. 1. 9. 1. 4. 0. 4. 4. 4. 4. 9.] [ 1. 1. 4. 1. 4. 4. 1. 1. 4. 0. 9. 4. 9. 4.] [ 1. 4. 1. 4. 1. 1. 9. 9. 4. 9. 0. 4. 1. 4.] [ 1. 9. 4. 1. 9. 1. 1. 4. 4. 4. 4. 0. 4. 1.] [ 1. 4. 1. 4. 4. 4. 9. 4. 4. 9. 1. 4. 0. 1.] [ 1. 9. 4. 1. 9. 1. 4. 1. 9. 4. 4. 1. 1. 0.]] For this matrix, I'm getting this with numpy: 2774532095.971 But I have a feeling I'm exceeding the capacity of floats here. Does anyone have an idea for how to treat this? Is it absurd to think I could get a determinant of this matrix? Is there a python package that could help me? Many thanks for any answers. James -- http://mail.python.org/mailman/listinfo/python-list Are you sure NumPy return float results. As far as I know, it returns doubles (about 16 digits) -- Home is not where you are born, but where your heart finds peace - Tommy Nordgren, The dying old crone [EMAIL PROTECTED] -- http://mail.python.org/mailman/listinfo/python-list
Re: Determinant of Large Matrix
James Stroud je napisao/la: Hello All, I'm using numpy to calculate determinants of matrices that look like this (13x13): [[ 0. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1.] [ 1. 0. 1. 4. 1. 9. 4. 4. 1. 1. 4. 9. 4. 9.] [ 1. 1. 0. 1. 4. 4. 9. 9. 4. 4. 1. 4. 1. 4.] [ 1. 4. 1. 0. 9. 1. 4. 4. 9. 1. 4. 1. 4. 1.] [ 1. 1. 4. 9. 0. 4. 4. 4. 1. 4. 1. 9. 4. 9.] [ 1. 9. 4. 1. 4. 0. 4. 4. 9. 4. 1. 1. 4. 1.] [ 1. 4. 9. 4. 4. 4. 0. 1. 1. 1. 9. 1. 9. 4.] [ 1. 4. 9. 4. 4. 4. 1. 0. 4. 1. 9. 4. 4. 1.] [ 1. 1. 4. 9. 1. 9. 1. 4. 0. 4. 4. 4. 4. 9.] [ 1. 1. 4. 1. 4. 4. 1. 1. 4. 0. 9. 4. 9. 4.] [ 1. 4. 1. 4. 1. 1. 9. 9. 4. 9. 0. 4. 1. 4.] [ 1. 9. 4. 1. 9. 1. 1. 4. 4. 4. 4. 0. 4. 1.] [ 1. 4. 1. 4. 4. 4. 9. 4. 4. 9. 1. 4. 0. 1.] [ 1. 9. 4. 1. 9. 1. 4. 1. 9. 4. 4. 1. 1. 0.]] For this matrix, I'm getting this with numpy: 2774532095.971 But I have a feeling I'm exceeding the capacity of floats here. Does anyone have an idea for how to treat this? Is it absurd to think I could get a determinant of this matrix? Is there a python package that could help me? Many thanks for any answers. James have you tried using matlab to verify the result? matlab is very fast and can work with large matrices, so this should be no problem for it... -- http://mail.python.org/mailman/listinfo/python-list
Re: Determinant of Large Matrix
On Jun 6, 6:47 am, Tommy Nordgren [EMAIL PROTECTED] wrote: On 6 jun 2007, at 13.10, James Stroud wrote: Hello All, I'm using numpy to calculate determinants of matrices that look like this (13x13): [[ 0. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1.] [ 1. 0. 1. 4. 1. 9. 4. 4. 1. 1. 4. 9. 4. 9.] [ 1. 1. 0. 1. 4. 4. 9. 9. 4. 4. 1. 4. 1. 4.] [ 1. 4. 1. 0. 9. 1. 4. 4. 9. 1. 4. 1. 4. 1.] [ 1. 1. 4. 9. 0. 4. 4. 4. 1. 4. 1. 9. 4. 9.] [ 1. 9. 4. 1. 4. 0. 4. 4. 9. 4. 1. 1. 4. 1.] [ 1. 4. 9. 4. 4. 4. 0. 1. 1. 1. 9. 1. 9. 4.] [ 1. 4. 9. 4. 4. 4. 1. 0. 4. 1. 9. 4. 4. 1.] [ 1. 1. 4. 9. 1. 9. 1. 4. 0. 4. 4. 4. 4. 9.] [ 1. 1. 4. 1. 4. 4. 1. 1. 4. 0. 9. 4. 9. 4.] [ 1. 4. 1. 4. 1. 1. 9. 9. 4. 9. 0. 4. 1. 4.] [ 1. 9. 4. 1. 9. 1. 1. 4. 4. 4. 4. 0. 4. 1.] [ 1. 4. 1. 4. 4. 4. 9. 4. 4. 9. 1. 4. 0. 1.] [ 1. 9. 4. 1. 9. 1. 4. 1. 9. 4. 4. 1. 1. 0.]] For this matrix, I'm getting this with numpy: 2774532095.971 But I have a feeling I'm exceeding the capacity of floats here. Does anyone have an idea for how to treat this? Is it absurd to think I could get a determinant of this matrix? Is there a python package that could help me? Many thanks for any answers. James -- http://mail.python.org/mailman/listinfo/python-list Are you sure NumPy return float results. As far as I know, it returns doubles I don't know about NumPy, but in general, a python float is a double: http://docs.python.org/lib/typesnumeric.html Floating point numbers are implemented using double in C. All bets on their precision are off unless you happen to know the machine you are working with. (about 16 digits) -- Home is not where you are born, but where your heart finds peace - Tommy Nordgren, The dying old crone [EMAIL PROTECTED] -- http://mail.python.org/mailman/listinfo/python-list
Re: Determinant of Large Matrix
James Stroud wrote: Hello All, I'm using numpy to calculate determinants of matrices that look like this (13x13): [[ 0. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1.] [snip] But I have a feeling I'm exceeding the capacity of floats here. Does anyone have an idea for how to treat this? Is it absurd to think I could get a determinant of this matrix? Is there a python package that could help me? Many thanks for any answers. James in order to verify that this result is correct, you could get the eigenvector and eigenvalues of this matrix (for example with numpy.linalg.eig) and check that they make sense; the determinant simply is the product of all the eigenvalues. gd luck -- http://mail.python.org/mailman/listinfo/python-list
Re: Determinant of Large Matrix
James Stroud wrote: Hello All, I'm using numpy to calculate determinants of matrices that look like this (13x13): [[ 0. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1.] [ 1. 0. 1. 4. 1. 9. 4. 4. 1. 1. 4. 9. 4. 9.] [ 1. 1. 0. 1. 4. 4. 9. 9. 4. 4. 1. 4. 1. 4.] [ 1. 4. 1. 0. 9. 1. 4. 4. 9. 1. 4. 1. 4. 1.] [ 1. 1. 4. 9. 0. 4. 4. 4. 1. 4. 1. 9. 4. 9.] [ 1. 9. 4. 1. 4. 0. 4. 4. 9. 4. 1. 1. 4. 1.] [ 1. 4. 9. 4. 4. 4. 0. 1. 1. 1. 9. 1. 9. 4.] [ 1. 4. 9. 4. 4. 4. 1. 0. 4. 1. 9. 4. 4. 1.] [ 1. 1. 4. 9. 1. 9. 1. 4. 0. 4. 4. 4. 4. 9.] [ 1. 1. 4. 1. 4. 4. 1. 1. 4. 0. 9. 4. 9. 4.] [ 1. 4. 1. 4. 1. 1. 9. 9. 4. 9. 0. 4. 1. 4.] [ 1. 9. 4. 1. 9. 1. 1. 4. 4. 4. 4. 0. 4. 1.] [ 1. 4. 1. 4. 4. 4. 9. 4. 4. 9. 1. 4. 0. 1.] [ 1. 9. 4. 1. 9. 1. 4. 1. 9. 4. 4. 1. 1. 0.]] For this matrix, I'm getting this with numpy: 2774532095.971 But I have a feeling I'm exceeding the capacity of floats here. It's not that you're exceeding the capacity of float64 numbers, it's just that there are floating point calculations taking place. The way the determinant is calculated is by doing an LU decomposition and then multiplying down the diagonal. Although all of your entries started as integers, floating point error does accumulate. The answer that you got is within finfo(float64).eps of relative error of the actual answer. Does anyone have an idea for how to treat this? Is it absurd to think I could get a determinant of this matrix? Is there a python package that could help me? If all of your matrices are going to be integers, doing the determinant-by-minors calculations yourself is probably easy enough to code and will retain complete precision. http://mathworld.wolfram.com/DeterminantExpansionbyMinors.html -- Robert Kern I have come to believe that the whole world is an enigma, a harmless enigma that is made terrible by our own mad attempt to interpret it as though it had an underlying truth. -- Umberto Eco -- http://mail.python.org/mailman/listinfo/python-list
Re: Determinant of Large Matrix
James Stroud wrote: I'm using numpy to calculate determinants of matrices that look like this (13x13): [[ 0. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1.] [ 1. 0. 1. 4. 1. 9. 4. 4. 1. 1. 4. 9. 4. 9.] [ 1. 1. 0. 1. 4. 4. 9. 9. 4. 4. 1. 4. 1. 4.] [ 1. 4. 1. 0. 9. 1. 4. 4. 9. 1. 4. 1. 4. 1.] [ 1. 1. 4. 9. 0. 4. 4. 4. 1. 4. 1. 9. 4. 9.] [ 1. 9. 4. 1. 4. 0. 4. 4. 9. 4. 1. 1. 4. 1.] [ 1. 4. 9. 4. 4. 4. 0. 1. 1. 1. 9. 1. 9. 4.] [ 1. 4. 9. 4. 4. 4. 1. 0. 4. 1. 9. 4. 4. 1.] [ 1. 1. 4. 9. 1. 9. 1. 4. 0. 4. 4. 4. 4. 9.] [ 1. 1. 4. 1. 4. 4. 1. 1. 4. 0. 9. 4. 9. 4.] [ 1. 4. 1. 4. 1. 1. 9. 9. 4. 9. 0. 4. 1. 4.] [ 1. 9. 4. 1. 9. 1. 1. 4. 4. 4. 4. 0. 4. 1.] [ 1. 4. 1. 4. 4. 4. 9. 4. 4. 9. 1. 4. 0. 1.] [ 1. 9. 4. 1. 9. 1. 4. 1. 9. 4. 4. 1. 1. 0.]] For this matrix, I'm getting this with numpy: 2774532095.971 But I have a feeling I'm exceeding the capacity of floats here. Does anyone have an idea for how to treat this? Is it absurd to think I could get a determinant of this matrix? Is there a python package that could help me? Here's some anecdotal evidence that your result may be correct: import operator m = eval([[ 0. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1.] [ 1. 0. 1. 4. 1. 9. 4. 4. 1. 1. 4. 9. 4. 9.] [ 1. 1. 0. 1. 4. 4. 9. 9. 4. 4. 1. 4. 1. 4.] [ 1. 4. 1. 0. 9. 1. 4. 4. 9. 1. 4. 1. 4. 1.] [ 1. 1. 4. 9. 0. 4. 4. 4. 1. 4. 1. 9. 4. 9.] [ 1. 9. 4. 1. 4. 0. 4. 4. 9. 4. 1. 1. 4. 1.] [ 1. 4. 9. 4. 4. 4. 0. 1. 1. 1. 9. 1. 9. 4.] [ 1. 4. 9. 4. 4. 4. 1. 0. 4. 1. 9. 4. 4. 1.] [ 1. 1. 4. 9. 1. 9. 1. 4. 0. 4. 4. 4. 4. 9.] [ 1. 1. 4. 1. 4. 4. 1. 1. 4. 0. 9. 4. 9. 4.] [ 1. 4. 1. 4. 1. 1. 9. 9. 4. 9. 0. 4. 1. 4.] [ 1. 9. 4. 1. 9. 1. 1. 4. 4. 4. 4. 0. 4. 1.] [ 1. 4. 1. 4. 4. 4. 9. 4. 4. 9. 1. 4. 0. 1.] [ 1. 9. 4. 1. 9. 1. 4. 1. 9. 4. 4. 1. 1. 0.]].replace(., .,).replace(], ],))[0] M = [[int(x) for x in row] for row in m] def subdet(m, rowindex): return [row[1:] for index, row in enumerate(m) if index != rowindex] def det(m): if len(m) == 1: return m[0][0] sign = 1 sigma = 0 for index, row in enumerate(m): x = row[0] if x: sigma += sign * x * det(subdet(m, index)) sign = -sign return sigma def common_multiple(items): items = set(items) items.discard(0) if items: return reduce(operator.mul, items) else: return 0 def det3(m, switch_algo=8): p = 1 q = 1 while 1: if len(m) == switch_algo: a, b = divmod(p*det(m), q) assert b == 0 return a cm = common_multiple(row[0] for row in m) if cm == 0: return 0 sign = 1 e = enumerate(m) for first_index, first_row in e: if first_row[0]: f = cm // first_row[0] assert (cm % first_row[0]) == 0 p *= sign * cm q *= f first_row[:] = [f*x for x in first_row[1:]] break first_row[:] = first_row[1:] sign = -sign for index, row in e: if row[0]: f = cm // row[0] assert (cm % row[0]) == 0 q *= f row[:] = [f*x - fx for x, fx in zip(row[1:], first_row)] else: row[:] = row[1:] del m[first_index] if __name__ == __main__: import pprint pprint.pprint(M) result = det3(M) assert result == 2774532096 print det(M) =, result As I use only integers, any errors should be algorithmic rather than caused by rounding. Peter -- http://mail.python.org/mailman/listinfo/python-list
Re: Determinant of Large Matrix
James Stroud wrote: For this matrix, I'm getting this with numpy: 2774532095.971 But I have a feeling I'm exceeding the capacity of floats here. Does anyone have an idea for how to treat this? Not if you don't state your requirements more precisely. E. g. what precision do you need? Is it absurd to think I could get a determinant of this matrix? Absolutely not. Is there a python package that could help me? Help doing what? BTW, scilab says this: --det(A) ans = 2.774532096E+09 --ans-2774532095.971 ans = 2.8610229492188E-06 -- The error is 15 magnitudes below your result -- what do you want more? :) Regards, Björn P.S.: print 2774532095.971 2774532096.0 -- BOFH excuse #118: the router thinks its a printer. -- http://mail.python.org/mailman/listinfo/python-list
Re: Determinant of Large Matrix
On Wed, 06 Jun 2007 04:10:43 -0700, James Stroud wrote: Hello All, I'm using numpy to calculate determinants of matrices that look like this (13x13): [snip matrix] For this matrix, I'm getting this with numpy: 2774532095.971 But I have a feeling I'm exceeding the capacity of floats here. Does anyone have an idea for how to treat this? Is it absurd to think I could get a determinant of this matrix? Is there a python package that could help me? Is there a particular reason you think there is a problem? The determinant given is pretty close to the integer 2774532096. Assuming that is the correct value, the difference between: 2.774532096000e9 and 2.774532095971e9 gives a relative error of 1.0311731312618234e-13 percent. How much precision were you after? :-) I suspect that if there is a problem with the matrix, it is less likely to be because of the size of floats and more likely that the matrix is ill-conditioned. I don't know if numpy will calculate the condition number of the matrix, or estimate it. If it does, do so -- a large condition number == trouble. http://en.wikipedia.org/wiki/Condition_number Another way to see if the matrix is ill-conditioned is to make a small perturbation to it (say, change two or three of the entries by 0.0001 or so), then calculate the determinate. If the result is radically different, then the matrix is probably ill-conditioned and there is likely no help for you except numerical black magic and/or using a different matrix. -- Steven. -- http://mail.python.org/mailman/listinfo/python-list
Re: Determinant of Large Matrix
Hello, Thank you to those who responded for your answers. They were very helpful and I'm confident now that numpy is calculating accurate determinants for these matrices. But I think I need to restate my problem a little as suggested by some becuase I'm still bewildered. First, here is the relevant part of my code: # start of code def main(word, repeats, trials): encodings = encode(word, CODE) random.shuffle(encodings) values = [] pb = PB(tk.Tk()) for i in xrange(trials): if i % 100 == 0: pb.update(float(i)/trials) random.seed() encodings = random.sample(encodings, repeats) distmat = build_distmat(encodings) print print distmat values.append(det(distmat)) values.sort() # print values print set(values) # end of code Here are some notes about the code: 1. PB is a progress bar because this takes a while for a lot of trials. 2. encodings is a complete list equivalent encodings of word. 3. distmat is a representation of the distances between the sample as integers for purposes of calculating the content of the hyperspace defined by this pairwise distance matrix (Cayley-Menger determinant). 4. det is numpy.linalg.det I get fairly unintuitive results, because if I run it once for 10 (or 1000 or 100,000) the set of values are all essentially the same: set([-733163520.5591, -733163520.4566, -733163519.8546, -733163520.2789, -733163519.9189, -733163519.783, -733163519.758, -733163520.2348, -733163519.6936, -733163520.4542]) If I run it again on 10 (or 1000) the set is basically homogenous but now of different values (terribly confusing): set([12048175104.1, 12048175104.15, 12048175104.46, 12048175103.94, 12048175104.23, 12048175103.81, 12048175103.98, 12048175103.9]) How could this be? This holds for 10 trials or 10 trials. Below is the output from the above runs. Is this a problem with the rng? All of these matrices look different. I'm certain this can't happen by chance. Thank you again. James euler 70% ./simplex.py [[ 0. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1.] [ 1. 0. 4. 9. 9. 4. 4. 9. 9. 1. 9. 16. 9. 9.] [ 1. 4. 0. 9. 9. 4. 4. 16. 16. 4. 9. 9. 1. 4.] [ 1. 9. 9. 0. 9. 9. 16. 4. 4. 16. 9. 1. 4. 4.] [ 1. 9. 9. 9. 0. 16. 4. 4. 9. 9. 9. 9. 4. 9.] [ 1. 4. 4. 9. 16. 0. 9. 4. 4. 4. 4. 9. 9. 4.] [ 1. 4. 4. 16. 4. 9. 0. 16. 9. 1. 9. 9. 9. 9.] [ 1. 9. 16. 4. 4. 4. 16. 0. 1. 9. 4. 4. 9. 4.] [ 1. 9. 16. 4. 9. 4. 9. 1. 0. 4. 9. 1. 9. 9.] [ 1. 1. 4. 16. 9. 4. 1. 9. 4. 0. 9. 9. 9. 9.] [ 1. 9. 9. 9. 9. 4. 9. 4. 9. 9. 0. 9. 16. 1.] [ 1. 16. 9. 1. 9. 9. 9. 4. 1. 9. 9. 0. 4. 4.] [ 1. 9. 1. 4. 4. 9. 9. 9. 9. 9. 16. 4. 0. 9.] [ 1. 9. 4. 4. 9. 4. 9. 4. 9. 9. 1. 4. 9. 0.]] [[ 0. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1.] [ 1. 0. 4. 16. 4. 9. 16. 4. 4. 9. 9. 1. 9. 9.] [ 1. 4. 0. 9. 9. 1. 9. 9. 9. 16. 9. 4. 4. 9.] [ 1. 16. 9. 0. 4. 4. 1. 9. 9. 9. 4. 9. 9. 1.] [ 1. 4. 9. 4. 0. 16. 9. 9. 1. 9. 4. 1. 9. 9.] [ 1. 9. 1. 4. 16. 0. 4. 4. 16. 9. 4. 9. 9. 4.] [ 1. 16. 9. 1. 9. 4. 0. 9. 16. 9. 9. 9. 4. 4.] [ 1. 4. 9. 9. 9. 4. 9. 0. 4. 1. 4. 4. 9. 9.] [ 1. 4. 9. 9. 1. 16. 16. 4. 0. 4. 4. 4. 4. 9.] [ 1. 9. 16. 9. 9. 9. 9. 1. 4. 0. 4. 9. 9. 9.] [ 1. 9. 9. 4. 4. 4. 9. 4. 4. 4. 0. 9. 16. 4.] [ 1. 1. 4. 9. 1. 9. 9. 4. 4. 9. 9. 0. 9. 16.] [ 1. 9. 4. 9. 9. 9. 4. 9. 4. 9. 16. 9. 0. 9.] [ 1. 9. 9. 1. 9. 4. 4. 9. 9. 9. 4. 16. 9. 0.]] [[ 0. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1.] [ 1. 0. 9. 9. 4. 9. 4. 9. 9. 9. 9. 9. 16. 4.] [ 1. 9. 0. 1. 9. 4. 9. 4. 16. 9. 9. 9. 4. 1.] [ 1. 9. 1. 0. 9. 4. 9. 9. 9. 9. 16. 9. 4. 4.] [ 1. 4. 9. 9. 0. 1. 9. 9. 4. 9. 4. 16. 9. 9.] [ 1. 9. 4. 4. 1. 0. 16. 16. 9. 4. 9. 9. 4. 4.] [ 1. 4. 9. 9. 9. 16. 0. 1. 4. 4. 4. 4. 4. 16.] [ 1. 9. 4. 9. 9. 16. 1. 0. 4. 9. 1. 9. 4. 9.] [ 1. 9. 16. 9. 4. 9. 4. 4. 0. 4. 1. 9. 9. 16.] [ 1. 9. 9. 9. 9. 4. 4. 9. 4. 0. 4. 1. 4. 9.] [ 1. 9. 9. 16. 4. 9. 4. 1. 1. 4. 0. 9. 9.
Re: Determinant of Large Matrix
Steven D'Aprano wrote: [Valuable Response] Thank you Steven for your helpful comments. Please see my reply to Bjoern Schliessmann where I have restated my problem. James -- http://mail.python.org/mailman/listinfo/python-list
Re: Determinant of Large Matrix
James Stroud wrote: [pointless stuff] OK. Nevermind. I'm rebinding encodings and so taking a sample from the sample and thus getting the sample back. Terribly sorry. James -- http://mail.python.org/mailman/listinfo/python-list
