[seqfan] Re: The Sequencer OEIS survey

[Philipp Emanuel Weidmann]

Thanks to all for the feedback.

Out of 14 conjectures on the page, 3 have now been proved and 2
disproved. I have added the proofs and disproofs to the page along with
links to the original Seqfan mails containing them.

As noted by several people (and mentioned on the page itself) the output
does not take OEIS offsets into account. The reason is that offset data
is not available for mass download (only sequence titles and elements
are) and thus not ready for parsing by software. IMO the data dumps OEIS
provides could do with some improvements: Rather than two separate text
files, it would be much preferable to have a relational database dump
containing all fields for all entries. Wikipedia has offered such dumps
for years (https://en.wikipedia.org/wiki/Wikipedia:Database_download)
and they are regularly used in academic endeavors like natural language
processing and data compression research. As data analysis technologies
advance, the OEIS' data could be the fuel of a lot of interesting ML
research if it were provided in an accessible format.

I intend to keep the URL
http://worldwidemann.com/the-sequencer-oeis-survey/ stable as requested.
Additionally, anchors for individual sequence entries are available via
the syntax
http://worldwidemann.com/the-sequencer-oeis-survey/#a056084httpoeisorga056084 etc.

Best regards
Philipp Emanuel Weidmann






On Sun, 2015-04-12 at 12:28 +0200, Philipp Emanuel Weidmann wrote:
> Back in February I announced a project to scan all OEIS sequences using
> the Sequencer system (https://github.com/p-e-w/sequencer) in order to
> identify new closed-form expressions for the sequence terms.
> 
> The search has now concluded and the results are available at
> http://worldwidemann.com/the-sequencer-oeis-survey/
> 
> Please note that only sequences without a "formula" field were scanned
> as explained in the article.
> 
> Best regards
> Philipp Emanuel Weidmann
> 
> 
> 
> _______________________________________________
> 
> Seqfan Mailing list - http://list.seqfan.eu/