Re: Pari programs on oeis, accessing internal calculatios

[Jamie Morken <jmorken@shaw.ca>]

Hi,

>what if the program does a(n)=(n+1)!-n! instead ?

I think the equal length sequences would be:

n+1, (n+1)!, n!, (n+1)!-n!.

>Let's say a(n) = sum(k=0,n,binomial(n,k)^3)

That is a good example, the simplest equal length sequences might be:

sum(k=0,n,binomial(n,k)^3), sum(k=0,n,binomial(n,k)), it is a bit arbitrary

to decide how to construct the equal length sequences but whatever method

is easiest to implement would work as a proof of concept.

It is a statistical approach to find relationships between A numbers,

so the most straightforward and simple implementation might be good

enough.

cheers,

Jamie

----- Original Message -----
From: Bill Allombert <Bill.Allombert@math.u-bordeaux.fr>
To: pari-dev@pari.math.u-bordeaux.fr
Sent: Wed, 06 Nov 2019 16:05:51 -0700 (MST)
Subject: Re: Pari programs on oeis, accessing internal calculatios

On Sun, Nov 03, 2019 at 05:07:58PM -0700, Jamie Morken wrote:
> Hi,For an example, since it is a general idea and not pari
> specific,for the sequence a(n)=n*n! for 1<n<10, there are three
> sequencesthat have equal lengths, ie, n, n!, n*n!

what if the program does a(n)=(n+1)!-n! instead ?

>For any pari program
> on OEIS, I was wondering if it could be possibleto break down the pari
> program to access the sequences that areused to create the overall
> a(n) sequence, including numerators anddenominator sequences of
> rational numbers.The main goal I was considering is to find as many
> sub-sequencesfor each oeis sequence as possible, and then build a
> graph data structurethat shows links between different oeis sequences,
> based on sharedsubsequences.

Let's say
a(n) = sum(k=0,n,binomial(n,k)^3)

what should that give ?

Cheers,
Bill