|
|
Post by coolrunning on Jul 20, 2020 14:41:20 GMT
Thanks for the input
My analysis programs are broken so an extra thanks for how they could be made to work and where they cannot be made to work.
Hi. Sorry for the delay - been in hospital and am now in enforced total relaxation, so it's been a few days . . . ... Hope this helps. Wow - comprehensive reply.
I just do some awful python programming. I now have to go back and fix the code in several places. But I lack the inspiration.
|
|
parisingoc
Member of DD Central
Posts: 87 
Likes: 25
|
Post by parisingoc on Jul 21, 2020 10:19:43 GMT
. . . . But I lack the inspiration.
I know that feeling  When I did my Computing Science Degree over 40 Years ago, I did all my programming in 370 Assembler - effectively machine code - and I loved it. These days, I just don't have the contiguous number of hours needed to concentrate the mind enough! (Yes, I am retired . . .). I find the visual nature of Rapidminer focusses me on the output. Yes it is a pain remembering or having to learn enough of the theory to be able to know what function might do what I think I want to do, but the speed of the product means I can be checking as I go. As a simple instance, simply dragging a "Read CSV" function onto the pallette and connecting the output of the function to the output tab of the pallette and selecting "Metadata View" reveals all sorts of facts about the CSV file being read including numbers of individual attribute values, missings, range, median values, counts per value of polynomials, etc, etc. I find just this basic analysis tells me which attributes can be relied upon and which are pretty much garbage. (It takes about 25 seconds to read the full Bondora Dataset). This then spurs me on to redefining the attribute type on the more important attributes so I can then use more functions on those attributes to get what I think I want. Perhaps the weakest item I have found in Rapidminer is a truth I learnt all those years ago which is that working with dates, times and all formatting combinations of the 2 is still a bummer and fraught with all the same problems that existed when I first learned to program (which I suddenly realise is now more than 50 years ago  !).
|
|
!).