Prev: reneerasmussen
Next: help needed in ocr(urdu version)
From: us on 26 Feb 2010 08:57
singlepoint <singlepoint(a)gmail.com> wrote in message <12ddc17d-698d-4f8c-a54d-433148f84f48(a)l24g2000prh.googlegroups.com>...
> I recently noticed this when I saved my some of my data in sparse
> matrix form and then I saved the transpose of this sparse matrix and I
> noticed a size reduction. Any ideas guys whats going on under the
> hood?
>
> Regards

here, this does not happen; can you show an example(?)...

us
From: singlepoint on 26 Feb 2010 09:11
Hey,
Lemme tell you the exact scenario. I have a long dataset and I have
converted it to (i,j,v) format. Example,
say 1 2 3 4 5 6 is one row of my dataset
I converted it to i,j,v format as follows

1 1 1
1 2 1
1 3 1
1 4 1
1 5 1
1 5 1

where first column is the row number (since I am considering mine is
the 1st row so is a 1 in first column everywhere) second column is the
actual value in the row and third column is whether this item is
present or not (1/yes, 0/no).
Now, since there were some 20k rows in the dataset when I converted
that dataset to this format It became even bigger. So, this is where I
noticed the difference in sizes. When I created a sparse matrix of my
converted dataset and saved it, it was around 2mb but when I took the
transpose and saved again, size came out to be 1.9mb.
I am sorry but I don't know how to show you guys an example of that
magnitude.
 | 
Pages: 1
Prev: reneerasmussen
Next: help needed in ocr(urdu version)