Nima Sarajpoor

> * Is there any reason why [`k` is being passed to `_compute_diagonal`](https://github.com/TDAmeritrade/stumpy/blob/main/stumpy/stump.py#L49)? I don't see it being used anywhere. So, I think I forgot to remove it probably. Previously,...

@alvii147 > Honestly it's so weird, somehow while `stump` has gotten faster, `stump_tiles` algorithm has gotten super slow. Now it takes for us to get at least `250k` time series...

Btw, regarding the hypothesis testing, the assumption is that the computing time samples are normally distributed.

**Idea 4:** As discussed in VALMOD paper (see #585), it is possible to find lower bound (LB) of distance between two subsequences of length `m+1` according to the distance of...

@seanlaw > @NimaSarajpoor If you comment on this then I can assign it to you Sure. I am going to work on this after we develop `top-k matrix profile` feature...

> I'm guessing that 99% of the time users will only care about `k=1` (and therefore they only care about 1D output). What do you think? I think 2D output...

> then maybe we should make check to see `if P.shape[1] == 1 and k == 1:` and then return a 1D array. Otherwise, return 2D. Something like that? Yeah...that...

@seanlaw @udf2457 I tried to implement the `Table I` and `Table II` of the paper, referred to as `DRAG` algorithm by the authors. The `DRAG` algorithm will be used later...

@seanlaw @udf2457 I tried the code with and without `break`. If we don't use `break`, the number of discords (before being pruned in the second phase of the algorithm) is...

@seanlaw: Apparently, the authors didn't use Matrix Profile in their algorithms. In fact, if you read the first paragraph of the section II-C of the paper, it says they are...