Weblog
Weblog
Seite 32 von 96, 1912 Einträge insgesamt
NITLE and Maciej are indexing weblogs with a crawler to give themselves a decent base of data for search testing.
There's also a very quick search engine for Kryogenix based on Maciej's Search::ContextGraph code, which He's ported to Python.
Despamming and other speculations
with some background explanations. LESEN
can be seen
here. Generated by the simple
zcat -c access_log-*.gz | grep 63.148.99 >cyv_traumwind.txt invocation (my logs get zipped each night)
to your right you see my shiny new 'brute force keyword hunter' thingie. Tajes a list of keywords, scans each and every post I ever made here, and returns links to those that have matches. Rather fast it is too. Pure regex power, no caching, no database and no sophistication. Took me 2+ hours to write and will be a basic example for that essay I was linking to some time ago.
In November, 2002, the GraphicsMagick project was formed as a branch from ImageMagick 5.5.2. The project was formed in order to develop a more "stable" form of ImageMagick which is suitable to be use as a base for other open source (and proprietary) applications. It was perceived that ImageMagick provided tremendous value, but also proved to be difficult to rely on as a stable application component since the ImageMagick interfaces often changed between releases, and there was no concept of shared library ABI stability. GraphicsMagick intends to resolve these issues while providing an appealing development environment to attract the best available developers.
esp the 'stable interfaces' part seems to be important, as I have to date been unable to install Perl's Image::Magick on my SuSE 7.2 webserver, try as hard as I did.
this looks intersting: GraphicsMagick Perl API
Note that this differs from the ImageMagick version of PerlMagick which uses the namespace Image::Magick. Any PerlMagick code written for the ImageMagick version of PerlMagick requires a global
substition of Image::Magick to Graphics::Magick
in order to work with the GraphicsMagick version.
Stealing is wrong. Misappropriating the property of another,
intellectual or otherwise is wrong. Yet companies with reputations as
shining as the Washington Post and Ford Motor Co., don't seem to mind
hiring a company to do, in their name, precisely what they so rightly
complain about
I can't really find answers to:
- What messages are considered when comparing for Junk? All the 'tagged as Junk' ones?
- What happens when I delete messages that are tagges as junk? does Moz 'forget' those examples?
- Is it better to just keep a cretain amount of example Junk around? Or should I save all Spam/Junk?
background on this: I now have an example corpus of about 1300+ Spam/Junk messages, and I noticed a degrade in detection accuracy. Actually, I had a large corpus of Junk mail I trained Moz on, and found it was overzealous. Having now marked a lot of messages as 'not junk' I see the exact oposite, it doesn't detect some obvious Junk at all...
So is it better to just have a rather small (200+) corpus of example Junk or what?
And then, is there a difference between messages not marked at all and messages marked not junk???
I had an important 'employee talk' today I was sweating my hand about for some time now.
It went smoothly and without surprises. Afterward I wonder why I made such a fuss about it. I didn't influence the outcome of it, and I couldn't have in any case. What was offered is reasonable, but puts the risk more on my account. Still if everything work halfway ok, I'll not suffer compared to now. And if itr goes well, I'll be better off.
Only question I'm asking myself: 'Why did I make such a fuss? Why am I so negative?'
That's why I love PHP. It makes things so simple at times
following suit, I added my GeoURL to the right. I used
WhereOnEarth to find my coordinates...
next time I rant about people not resonding to email. some one please ask me this simple question: "Have you actually sent that mail, Martin?"
IHMC Concept Mapping Software
The IHMC Concept Map Software (CMapTools) from the University of West Florida empowers users to construct, navigate, share, and criticize knowledge models represented as Concept Maps. The toolkit is platform independent and network enabled, allowing the users to build, and collaborate during the construction of concept maps with colleagues anywhere on the network, as well as, share and navigate through others' models distributed on servers throughout Internet.
Search::ContextGraph.
Let the party begin! The Perl Module to do 'Contextual Network Graph' Searching... Oh yeesss!
Classifying stuff with Java / Couple of interesting packages for dealing with text. "Classifier4J
is a java library that provides an API for automatic classification of
text. The default (and only current) implementation of this API is a Bayesian classifier." and "EBayes
is a full engine for manipulation of Bayesian networks, aimed at applications that can afford minimal storage of executable code. The engine can handle arbitrary networks and can produce inferences, expected values and explanations (maximum a posteriori). The engine is based on the JavaBayes system, a complete system for creation and
manipulation of Bayesian networks. The EBayes engine has all the basic capabilities of the engine in JavaBayes, but has been quite optimized in a variety of ways."
Danny Ayers has been working on IdeaGraph, a Java application "for creating visual maps of ideas, that can work with web pages, documents and images".
My name is Maciej Ceglowski. I used to be an oil painter, and lately I've been a Perl programmer, but I'm trying to become an oil painter again.
One wish I can really feel also... Only I won't be going back for other reasons. And nothing much I can do about it.
on your right. But it seems to be not very well indexing this blog. The rest of my site is indexed anf found really fast, my essay is already in the index.
It appears the FOA book I linked some time ago is actually are very good resource on all this 'semantic search' stuff:
FOA: 6.5.2 Spreading activation search, which is an approach that is so similar to 'Contextual Network Graphs' that the authors of the latter now
feel like they 're-invented the wheel' ;)
Azeem Azhar has an excellent summary
of the two schools of thought in search engine design, and how they
pertain to Google's recently announced purchase of Applied Semantics.
must-read (for me). Isn't that just the topic I'm trying to grapple?
It is per design a-chronal:
Traumwind - a-chronical story telling
And why would that be a good ting(TM)? Because it can show me things I might already have forgotten about. While keyword search can show me things I just-quite remember, semantic similarity search can show me things, I have already forgotten, or never known or suspected.
neuere Einträge | ältere Einträge
alles Bild, Text und Tonmaterial ist © Martin Spernau,
Verwendung und Reproduktion erfordert die Zustimmung des Authors