Kommentare zu Rafazwonull vs Joe I/O

Kommentare von Melanie zu „Angela Merkel twittert nicht“ – Politik und Web 2.0

Melanie — Sat, 22 Aug 2009 12:01:50 +0000

Ich denke, dass liegt daran, dass langsam aber sicher sich „jeder“ über die bevorstehenden Wahlen informiert. Zudem wird dein Artikel in der google-Suche Angela Merkel+Web 2.0 als 3.! Artikel aufgelistet. Mein persönliches Interesse daran liegt in meinem Diplomarbeitsthema begründet.
Vielleicht hilft dir das ja ein bisschen weiter.

Kommentare von Rafa zu „Angela Merkel twittert nicht“ – Politik und Web 2.0

Rafa — Wed, 29 Jul 2009 21:07:15 +0000

Darf ich eine kurze Frage stellen? Im Juli 2008 riefen 25 Leser diesen Artikel auf. Im Juli 2009 waren es 270. Er ist ein Jahr(!) alt. War ich meiner Zeit mit diesem Post voraus? Ist er etwas Anderes als ihr euch vorgestellt habt, werte Besucher von Google und Co.? Oder ist irgendetwas total akut, was mir entgangen ist? Tipps und Kommentare sind hoch geschätzt.

Kommentare von Bringt es was, Twitter an StudiVz anzubinden? « Rafazwonull vs Joe I/O zu Offene APIs: Mein Facebook wird wissen, was mein Skype ihm sagt.

Bringt es was, Twitter an StudiVz anzubinden? « Rafazwonull vs Joe I/O — Wed, 08 Jul 2009 12:42:13 +0000

[...] Top-Beiträge "Angela Merkel twittert nicht" – Politik und Web 2.0Offene APIs: Mein Facebook wird wissen, was mein Skype ihm sagt. [...]

Kommentare von Marcel Knust zu Quality Model Development

Marcel Knust — Sat, 30 May 2009 21:49:08 +0000

Even if you have better models, I suppose that they would not be stable because of changing cultures and fast changing technologies of blog creation.

Kommentare von Quality Model Development « Rafazwonull vs Joe I/O zu Qualitätsmodelle und Data Mining in Blogs

Quality Model Development « Rafazwonull vs Joe I/O — Mon, 25 May 2009 21:35:34 +0000

[...] Top-Beiträge Findings on Quality Blog recognitionLocation Tagging: Neues zu Brightkite und LocatikQualitätsmodelle und Data Mining in Blogs [...]

Kommentare von Quality Model Development « Rafazwonull vs Joe I/O zu Data Cleansing and Quality Model Generation

Quality Model Development « Rafazwonull vs Joe I/O — Mon, 25 May 2009 21:26:13 +0000

[...] Data Cleansing and Quality Model Generation [...]

Kommentare von Rafa zu Findings on Quality Blog recognition

Rafa — Mon, 25 May 2009 18:50:02 +0000

Hey Stefan,
thanks for this contribution. Alas, I guess, this is the offside of data mining models: they are hard to visualize and J48 is maybe one of the most understandable ones. They are neither intuitive, nor undestandable and the models result purely from the data they are based upon.

Imagine one of the linear algorithms for example:
the algorithm recursively computes the weigth for each one of the attributes the decision models are based upon. Then, it takes each of the numeric values it has considered, multiplies it with the weigth and takes the sum of these for each instance. This is the computed numeric value of the class, the corelation of which is computed to the initial class values 1=good and 0=bad. This is at least what I have understood of what my supervisor has explained to me. Any suggestions on how to visualize that ? This is not even the toughest one: imagine vectors in multidimensional ~~rooms~~ spaces (Edit: Thanks Max!), where each of the features is a dimension… great considering 150 attributes…

Anyway, I will gladly post some examples on distribution graphs to supply you with some visual goodies, if you care for it! You are pointing at a very good issue indeed: I will post some data on J48 and explain to you what happened!

Kommentare von Stefan Martens zu Findings on Quality Blog recognition

Stefan Martens — Mon, 25 May 2009 18:03:59 +0000

Hey Rafi,

I could imagine there’re some people out there that would really like to see an example.

What do you think of posting an example blog with screenshots?

I reckon that’d make your work more visible to people that haven’t spent the last 6 months writing a thesis on that topic.

I’d really like to see that and am more than willing to contribute to your comments here AND leveraging the power of social media to lead some traffic here for a discussion.

Cheers,
Stefan

Kommentare von Rafa zu Findings on Quality Blog recognition

Rafa — Mon, 25 May 2009 16:21:19 +0000

Dear Mr -m- (or just „M“ as they call you in various agent thrillers),

thanks a lot for the compliments. I love you too. I would never get the idea of thinking badly of the great Russian people, their women or their kitchen. Neither of their blogosphere.

I will write some stuff on the language recognition as well. Actually, I would like continue the research on the heterogenity issues as well as I would like to dig deeper into text mining issues, to really get more hold of semantic quality. But this is something for a PHD or another thesis!

However, what we have here as a method and a tool, we could easily extend to other fields: finding people on the net who fit job descriptions, find places which fit peoples profiles, find guys which fit girls. Once you have understood the concept, this is really powerful. We should have an own company! Anybody knowing VCs to be trusted?

Kommentare von max zu Findings on Quality Blog recognition

max — Mon, 25 May 2009 08:13:24 +0000

Hey Mr. -r- (that sounds quite interesting )
I reckon that your last comment in the brackets was just an intension to include me in the conversation (I could call it an intrigue as well). Of course the russian blogs work the best, there is no doubt and I cannot recall a time in which people tried to doubt exactly this fact without loosing their head. So far to the historical part of our issue!
I am pretty impressed by the results and your conclusions seem to be informative and proven by your research. Of course you create new issues which you’ve already listed in your summary. I am looking forward to hear more about that! Will you actually continue with your research or you really want to tell me that you are not interested in finding out why the english blogs have that high heterogeneity?!

Take care!
- m -