Data Mining

Do any of you guys know Data mining? Also what kind; video, database, web, etc...I need some contacts to share information, etc.

 

Log in or Sign up to hide all adverts.

Come on! There's got to be at least ONE!

 

Log in or Sign up to hide all adverts.

I studied Theoritical Data Mining as part of ERP in subject form,that is why i didnt reply,because i know i wouldnt be able to give any practical tips.

bye!

 

Log in or Sign up to hide all adverts.

Perhaps if you're a bit more specific of what you're looking for. Data mining is a rather general concept, i.e. it is used to describe all search/compile/aggregate/report facilities with 1 or more data sources, although it is most commonly used to describe data collection processes in large databases.

Are you looking for general information pertaining to data mining? Specific algorithms? Ways of collecting disparate information? Or just a general group for data mining chit chat?

 

Essentially, I was stupid enough to branch out of engineering and pick a masters thesis in data mining. I am looking for individuals with knowledge that tanscends the general-- can be those currently pursruing degree in data mining, or have practical experience in the field. I use data mining loosely and one can liken it to 'knowledge discovery". Anyone working in AI and has used neural nets and like, is welcome; those that knowing mining for data in very large databases are also welcome, etc. In fact if you know of the quest group at IBM or more improtantly, if you know any others places for papers, etc. you are welcome.

 

The information I use to look at on datamining USED to be housed at www.botspot.com, However it seems that the site went under new management and became more orientated around getting money from people.

So I would suggest using www.archive.org to look at the old botspot to gain some information/links etc.

You might also want to look at information on "Spiders/Bots", since spiders/bots "datamine" when they get information on a website and then store it for search engines (i.e. google)

[A few Spiders/Bots]
http://www.grub.org
http://www.inktomi.com/slurp.html
http://fast.no/support/crawler.asp
http://www.googlebot.com/bot.html

 

Monsieur, Merci.

 

Try a university library. Reference section. No bots allowed.

 

KDnuggets

Another good site to look at is:
http://www.kdnuggets.com/

Its a directory for sites related to data mining. I don't know what you mean by "you decided to do a thesis on data mining". Do you have a specific focus? Data mining is so huge I can't imagine you planning to do a summary evaluation for a thesis.

If you're going to specialize and you'd actually like to get the paper done I'd recommend looking at text mining. It is by far the most mature knowledge discovery application so you'll actually find some material written about it.

 

Anyone know the incremental EM algorthm developed by Stauffer and Grimson for detecting moving objects?

The paper appears in IEEE transactions on Pattern analysis and machine intelligence(PAM) 2000. I want the paper or the transactions for free...

Please Register or Log in to view the hidden image!

 

Already had it amongst all the crap... Thx anyway


voice...
More along the line of motion detection, etc using clusters.... spatiotemporal blocks...

Beautiful, importnat data to work with too..

 

That sounds really interesting fountainhead. I'd be really interested in reading whatever you're doing, I think improved vision is going to be the driving technology behind a massive new wave of robotics.

The hardware obviously isn't there yet for real time artificial vision, but it will be soon (15-20 years). What is your technique? Are you breaking each image into blocks (pixels) and assigning them values based on criteria you've developed? Then comparing the map you create from the values with previously created maps? Most facial recognition software works sorta in that way.

Why cluster analysis, why not neural nets?

The Fountainhead is one of my favorite books, although my favorite Rand book is "We the Living".

 

voice of reason,

Yea the kdnuggets site is pretty damn good-- free data you can experiment on; you cannot beat that...

That sounds really interesting fountainhead. I'd be really interested in reading whatever you're doing, I think improved vision is going to be the driving technology behind a massive new wave of robotics.
Let's hope I get a publication... But yea, you are right. For the past few years there has been a lot done in pattern recognition, motion detection, etc. Whether improved vision will be the driving technology behind robotics is however debatable. You should check out MIT's AI lab.

Are you breaking each image into blocks (pixels) and assigning them values based on criteria you've developed? Then comparing the map you create from the values with previously created maps? Most facial recognition software works sorta in that way.
Well, not everything I'm doing or thinking of doing is novel; hell most isn't... Also, we are using videos... Actually there is a tehnique the professor who I plan to work with/ am working with used whereby he uses '3-d blocks --- spatiotemporal blocks. In order words, you have multiple frames that are represented in vector form, the vector block decreased, and a variant of the EM algroithm by Stauffer that uses Gaussian distribution to group blocks as moving or not moving is used, then the blocks classified as moving are tracked.... I hope to use a modified version for my data.. well there is more to it.

Why cluster analysis, why not neural nets?
Data dependent

The Fountainhead is one of my favorite books, although my favorite Rand book is "We the Living".
Rand is Ok. If you will notice, my name is the fountainHED-- in derision....sorry. I do like the Wynand character in the book though, I must say. When it comes to western philosophy and the like, I prefer Kant above all.