Added Lab1 report #12
Conversation
There was a problem hiding this comment.
Code and report looks good. Try to expand the tests for future assignments.
Try to include a link to a Google Collab file (see template at the master branch) so that we also can run your code.
As Lab 2 is included in this PR, it will not get merged until Lab 2 is graded.
There was a problem hiding this comment.
Other than the comments in the review, for future reference: Your tests need to be more exhaustive. Small matrices are not enough as they do not necessarily show how accurate your implementations are. Instead create larger random matrices and compare your results to the ones gotten by numpy's implementations.
All changes mentioned in this review are needed to pass lab 2.
| " if (A[i, j] != 0): return False\n", | ||
| " return True\n", | ||
| "\n", | ||
| "def qrf_test():\n", |
There was a problem hiding this comment.
The norms to test against mentioned in the lab instructions are missing: ||Q^tq - I|| and ||QR-A||.
These need to be added to pass the lab. Also consider testing the accuracy of your implementation to the one of numpy's.
| "colab": {} | ||
| }, | ||
| "source": [ | ||
| "def ds(A, b):\n", |
There was a problem hiding this comment.
You are here using numpy directly, which is not allowed. This assignment (assignment 3) needs to utilize the algorithms mentioned in the theory, see lecture notes part 3 for details.
|
Lab 3 looks great, but lab 2 still needs to be updated: see the changes I requested above. |
| "cell_type": "markdown", | ||
| "source": [ | ||
| "### Monte Carlo quadrature over a unit interval ###\n", | ||
| "We check that the error converges towards zero as the number of samples increases. By inspection it is apparent that the convergence rate resembles 1/sqrt(samples)." |
There was a problem hiding this comment.
To make it clear that the convergence rate resembles 1/sqrt(n) you need to actually plot 1/sqrt(n) as well. Or a log-log plot. Comparison is a must for you to strengthen your claim.
No description provided.