I have a dataset

7000 items with a single value eg: 125.

72 different ways to look at each of these items. (eg: 10x125; 15x125; 25x125)

eg:

option 1: 10, 15, 25,

option 2: 10.2, 15.2, 25.2,

...

option 24: 12.2, 12.25, 13.2,

so I will end up with 72 sets of results for each of the 7000 items. (3 options x 24) x 7000 = 72 x 7000.

72 results.. rank them 1 to 7000 for each of the 72 results sets.

average(summed rank/divided by 72) = item overall score.

ie: in plain english from the 7000 items i have, I run 72 tests on each item (eg: multiplication), found the average score (as a ranking 1-7000) for each item over the 72 tests, and hey presto, i end up with a scale of what performs best in every scenario.

Is this a valid method, is there a better way? has my maths failed me? have i forgotten something obvious?

---

edit: maybe i could just run 72 tests on each item. Have a cumulative sum and divide the total by 72 to get a final rank.

Trouble is i dont want to spend the whole evening writing the code, I think the heat is getting to my brain.

Maybe i just needed to write these things down to get the ball rolling...

Edited March 23, 201016 yr by whiterussian

... ignore this thread, i think ive worked it out.. might just repost here in the future, if i get stuck again... it took me writing it down to an anonymous forum to clear my head!

I was really after some sort of graph with boundaries, upper/lower.. not sure of the name for this, i guess its deviations of some sort

but in the end, i will just crunch the numbers and see if what pops out looks right...

now to write the code for the 343,728,000 calculations - im going make my CPU suffer!

Edited March 23, 201016 yr by whiterussian

thank goodness for that

still want to know where i can source commercial actuarial tables

Edited March 23, 201016 yr by whiterussian

without context it's a little bit hard to come up with an alternate solution. that said feel free to use the forum for venting, I find that asking the right question and finding the answer are often the same thing...

...I find that asking the right question and finding the answer are often the same thing...

Very true! - in fact the above post i made is a load of rubbish - the data gave a linear result, so i didnt need to repeat the calculation umpteen times... now of course I'm stuck on the next equation which a maths student could probably write a single line equation for, but I will have to go the long route, and figure a way from the data backwards. I think i may be talking gibberish again. This thinking aloud thing is a wierd one - might be too much coffee!

I've had coders block for 4 days now. In fact, I'm only half way through the above problem.

The trouble I have with experimental projects, is that if i leave it a few days between working on it, I forget exactly what i was doing previously, its rather like juggling - Because you have to work out what you want to do before you design the script and the tables into which you are going to store the data. So in effect you have to do everything in your head, before you can start writing a single line of code.

Its very frustrating! (having two kids on school hols running around the house - doesn't really help that much either!! )

Edited March 24, 201016 yr by whiterussian

Pivot Table? Or have I misunderstood the objective?