http://www.talkstats.com/ New to TS? Stop in here and say hello! en Thu, 20 Jun 2013 06:43:20 GMT vBulletin 60 http://www.talkstats.com/images/misc/rss.png http://www.talkstats.com/ http://www.talkstats.com/showthread.php/44629-Hi-everyone?goto=newpost Wed, 19 Jun 2013 18:20:11 GMT Hi, My name is Patrick and I am currently writing my master thesis on antecedents of brand love. I´m studying marketing in Nijmegen, a small but... Hi,
My name is Patrick and I am currently writing my master thesis on antecedents of brand love. I´m studying marketing in Nijmegen, a small but wonderful city in the Netherlands nearby the German boarder. During data analyses I encountered some strange things, and after a search on some statistical solutions via google I became aware this forum. Overall, eager to learn more about statistics.
Kind regards,
Patrick ]]> New Member Introduction Patrick@RU http://www.talkstats.com/showthread.php/44629-Hi-everyone http://www.talkstats.com/showthread.php/44621-Hi-getting-oriented-)?goto=newpost Wed, 19 Jun 2013 10:49:38 GMT Hi there. I'm a lapsed computer scientist (currently building a house) who somehow never got around to studying statistics, either in school or university. Possibly one of the greatest mistakes of my life, as I begin to realise now!

While I'm doing lots of physical work, I have time to let my mind wander - which has raised some probably elementary questions which I can't answer, hence I'm here!!!

Perhaps somebody here could point me in the right direction, I don't even know which sub-forum to address :confused:

What's been on my mind...

It's come from thinking about network routing in computing. Imagine a (large) acyclic graph where data can take a random route from a source to a leaf via uniquely named nodes (like on the Internet from a server to a browser). Along the way the nodes influence the data according to a simple weighting function (e.g. perhaps this function adds latency, or reduce bandwidth, by a % value at each node).

Imagine that, as the data arrives at a leaf, the leaf knows the path through the graph which has been taken (something like a tracert) and has a value for the total influence exerted as it has passed through this network.

Would it be possible to (probabilistically) map out the weightings in each node - with sufficient data?

For example :

Imagine we have a bunch of paths and "output" values :
A>B>C>D>E = 0.9
B>G>C>E>A = 1.5
A>B>D>C>E = 0.9
J>U>C>E>E = 0.6
... (and many thousands, perhaps millions, of other recorded values)...

Is it possible to use statistics to uncover likely values for A..Z ???

Intuitively, it seems to me that (from the very little I know about it) something like Bays theorem, or Marcov Chains might be suitable, but I'm not sure about this...

Any pointers would be great, they'll give me something to think about while I'm bored to tears holding a trowel in one hand and a bucket of cement in the other!!!!

Cheers,
Ben ]]> New Member Introduction benautonomic http://www.talkstats.com/showthread.php/44621-Hi-getting-oriented-)