Some improvements to Tamas' work

[st...@nunez.org]

It's been kind of quiet around here and I thought I'd let everyone about some small improvements to Tamas Papps' libraries I've been working on over my COVID downtime. Mostly this is just documenting, integrating and putting them together as a package. I also added a special functions library, currently with all the normal functions implemented. The intention is to make this into a distributions library and remove the need to use librmath.

Comments from this group are especially welcome given the previous discussions over the years. In fact most of the direction was taken from those discussions. I'd love to get some feedback before announcing it 'officially' to a wider audience (the cat got out of the bag in March, before I was ready)

https://lisp-stat.dev/blog/2021/05/09/statistical-analysis-with-lisp-stat/

Regards,

    Steve

[Andrew Kirkpatrick]

This looks great thought I've not played with it yet.

The relationship to XLisp-Stat is not made clear in the overview, it would be good to see a paragraph on that or a section on the history of this code base.

Cheers,

Andy

[Steve Nunez]

Good point. Someone on the Hackernews site asked the same question. Except for def/undef/variables functions there is no XLisp-Stat code in this, it's either one of Tamas' libraries, or one of my own. I should probably remove that vestige of XLisp-Stat anyway, I don't think it's going to be useful going forward. But making the origin clear is a good idea.

--
You received this message because you are subscribed to a topic in the Google Groups "Common Lisp Statistics" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/lisp-stat/PPjfpAIBVH0/unsubscribe.
To unsubscribe from this group and all its topics, send an email to lisp-stat+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/lisp-stat/ce49d272-6fa4-4fc8-b018-111bfd9e4a7en%40googlegroups.com.

[Tony Rossini]

well, there are 2 lisp-stat's, one was the original, implemented with xlispdtat, and the other started from the incomplete 1990 Kyoto common lisp implementation, which i fixed up before realising that the overall approach was really dated and useless (no sense of dataframes/tables or groupings). 

i also started with Tamas' work, but personal life seöf destructed, and am only recently getting my life back (have had a few false starts).

key for me is dataframes, and Marco gave me what i was looking for - in my CLS repository, there is lots in the comments/docs per file to see what i was thinking of, more of a "novel data science DSL" which was kind of where R is with the tidyverse, but more decision oriented than math oriented....

You received this message because you are subscribed to the Google Groups "Common Lisp Statistics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lisp-stat+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/lisp-stat/CAMXtoqNWzeQvkw71ZnF2U6s5CjXpO6O_Y%3D7Pdu%2B3U44mp-N97A%40mail.gmail.com.

[Steve Nunez]

I also think getting a good data frame implementation is key to getting something off the ground. That's why most of the work was focused there and it's the top priority on the roadmap. Rho is interesting, and Marco writes good code. There's a lot that could be cherry-picked from there, like factors, which is what I'm working on now.

To view this discussion on the web visit https://groups.google.com/d/msgid/lisp-stat/CAA1MS4onkQsYB%3DCYZsgV2E9joofPtaQAi8RgQDaAUzV3nkOt7A%40mail.gmail.com.