Seems it returns "nan" with big matrix

[osoleole]

Seems it returns "nan" with big matrix >>100

1000x1000
Eigenvector squared calculated by...
numpy: 1.2163269368593035e-08
new-fancy: nan
/usr/local/lib/python3.6/dist-packages/numpy/core/fromnumeric.py:90: RuntimeWarning: overflow encountered in reduce
return ufunc.reduce(obj, axis, dtype, out, **passkwargs)
/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:22: RuntimeWarning: invalid value encountered in double_scalars

Seems this simple algorithm is not appropriate for real use. It's extremely slow and and isn't tuned.
Just nice math fact.

[ju2ez]

So you would say the algorithm is slower than the conventional one?

[osoleole]

Yes. Maybe it's just my naive implementation. There are many "for" cycles in it because we need to find many Minor's eigenvalues. I just didn't have a time to play with it more.

…

On Mon, 18 Nov 2019 at 14:17, Julian Hatzky ***@***.***> wrote: So you would say the algorithm is slower than the conventional one? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#2?email_source=notifications&email_token=ACBH273VJY626EBQS4MLFBLQUKPZDA5CNFSM4JOCVFLKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEEKSLTA#issuecomment-555034060>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACBH276BYFOUQCVWTS3O4JLQUKPZDANCNFSM4JOCVFLA> .