Link to the classroom item you are referring to: [https://www.coursera.org/learn/machine-learning-linear-algebra/lecture/k9Srg/the-gaussian-elimination-algorithm](Gaussian Elimination Algorithm)
at 3 minute mark.
In R3 - How did they get [0 3 - 2 2]?
If R3 - 4R1 is to be followed
It translates to -1 - 4(0.5) and results in -3. ( [0 3 -2 -3] )
I might be missing something, could i trouble anyone for an explanation?
@lucas.coutinho this does look like a mistake the R3 calculation for the last cell should -3 but has been presented as 2?? Can you have a look!!
I found this mistake on the video as well. It’s simple to verify, the video solution doesn’t work for the system presented earlier in the video. I believe the correct matrix shoul be as follows:
| 1 | 0 | 0 | 0 |
| 0 | 1 | 0 | -1 |
| 0 | 0 | 1 | 0 |
which results in a=0, b = -1, and c = 1
(mentor edit: the c value given here is incorrect. Please read the entire thread for more explanation)
Feedback: Respectfully, this not a minor mistake, and anyone working through this is likely to spend a lot of extra time trying to figure out what they did wrong. Advise you update the lecture or at least put in a disclaimer that some of the math is wrong. Note there’s also a mistake in “Row echelon form in general” but that one is much more minor than this one.
Adding: In previous thread I learned that updates to the videos are hard. Given that I suggested having a formal errata posted per week/video so people could first check that if they run into problems. A note in the course, such as a text overlay or pre-video note would be ideal, but the errata is lower effort and probably more approachable than searching the boards for many. Goal would be to reduce if not eliminate anyone getting so frustrated they just quit the course. Maybe just me, and somewhat paradoxically, the reason this video mistake is so frustrating is everything else is so well explained I assumed I was making an error, and spent hours trying to find it.
Just confirming what you’ve said – I spent several hours reworking this example, thinking I was doing something wrong. Even still, without confirmation I was getting the steps correct, I ended up with a different solution than you:
| 1 | 0 | 0 | 1 |
|---|---|---|---|
| 0 | 1 | 0 | 1 |
| 0 | 0 | 1 | 0 |
Without a corrected video, I have no idea where I went wrong.
The correct result is [0, -1, 0], as you posted previously but with a typo:
The error in the calculation of the right column of R3 propagates into all of the other rows, in complicated ways.
A support ticket has been submitted for this entire video.
thank you @TMosh! I don’t see a way to edit my post at this point, so hopefully people will read to the end of the thread.
Rather than break the continuity of the thread discussion (by retroactively making changes), I added a note to your earlier post.
