Lazy Programmer Forum

Thanks for your inquiry.

Slides may be requested at any time by contacting the instructor using the "contact" form at lazyprogrammer.me.

Best regards

Thanks for your inquiry. This is mostly a theoretical course in preparation for future courses, so one has to keep that in mind. Look at the final code lecture to see how we deal with it (empirical estimation). Or you can use the full Bayesian method in which the variance is assumed unknown and the ...

Good question. We left some symbols out there for brevity (this is normal). Recall that expected values are always with respect to a specific distribution. e.g. E_p [X] is different from E_q [X] In the Bayesian formulation we are interested in E[w | X, y] (w wrt the distribution p(w | X, y)), not wr...

Thanks for your inquiry.

Please note that the #1 source of student error is copying code incorrectly. Therefore, the first thing one should always do is cross-reference with the working code.

Let us know if that doesn't work, or if you found that the official code is not working.

Cheers!

Thanks for your inquiry. The reason is purely mathematical. If you want to learn the actual details of PCA, I'd recommend looking at the course "Unsupervised Deep Learning in Python" where this is discussed in detail. In the Finance course the section on PCA is only meant to look at the ma...

Thanks for your suggestion.

I consider pretty much everything in this lecture to be obvious (so much so, that it wasn't even added until years after the course already existed).

Which parts of it weren't obvious to you?

Thanks for your inquiry.

As you recall, MNIST is an image dataset. Some pixels have a constant value of 0.

In that case, the variance is 0. If the variance is 0 the PDF goes to infinity, which cannot be used for computations.

Thanks for your inquiry.

TF-IDF can be used on any non-standard symbols you like. In fact, in the course "Machine Learning: Natural Language Processing in Python (V2)" that is what we did.

Thanks for your inquiry.

It looks like you might be mixing up the "graph building phase" (symbolic variables) and the "inference phase" (numerical values).

The data links are given in the course (see the "where to get the code" lecture)