Reputation: 25
numpy.linalg.inv() doesn't work
I am trying to invert a 2-D matrix using the numpy.linalg.inv() function. However whenever I run it, it just seems to stop. Since I was using a matrix of dimensions 40989x52, I initially thought that this was an issue due to the so I tried waiting for it to process, but it just didn't proceed even after 30+mins. I later tried calling the inv() function with a 3x2 matrix but it surprisingly didn't work either.
I am using python3.6 with Anaconda libraries on 64-bit Ubuntu.
Could this be a library issue or is it an issue with my system?
Upvotes: 1
Views: 2424
Answers (1)
Reputation: 483
Whenever a matrix has more rows than columns (the system of equations has more equations than unknowns), the system could be overdetermined- in other words, there is no exact solution. In fact, you cannot take an inverse of a non-square matrix.
You might want to check the rank of the matrix.
However, practically, we end up using the pseudo-inverse for this scenario. You can look into using pinv.
https://docs.scipy.org/doc/numpy-1.14.0/reference/generated/numpy.linalg.pinv.html
Upvotes: 1
Related Questions
- np.linalg.inv() leads to array full of np.nan
- The function np.linalg.inv() doesn't work correctly
- inverse matrix np.linalg.inv in Python, gives weird results, why?
- What does scipy.linalg.lu return?
- numpy.linalg.inv returns inverse for a singular matrix
- Is the numpy.linalg.lstsq rcond parameter not working according to the description?
- Using inverse matrix function "inv" in python numpy.linalg
- numpy.linalg.lstsqr Errors
- Is numpy.linalg.inv() giving the correct matrix inverse? EDIT: Why does inv() gives numerical errors?
- python/numpy: problems with numpy linal.eig