Last week there was 5th Annual Postgraduate Conference 2009, Department of Computer Science & Software Engineering, University of Canterbury during 1-2 September 2009. 

The conference program was divided into 7 sessions: session 1-2 Honours, session 3-4 MSc, and session 5-6 Ph.D.

Here are Conference Sponsors:
Allied Telesis, Christchurch City Council, Dynamic Controls, Hewlett Packard, IBM, Jade, Microsoft, Orion Health, Schneider Electric, Tait Electronics, Telogis, and Trimble Navigation

Thus, the meeting of this week was postponed until Tuesday 8th, september.

Regarding to week7, the DNA sequence data were transformed to symbols via Sequitur. These symbols were preprocessing as a training sets (.arff) and also StringToWordVector filter function was chosen from preprocess in Weka. Afterthat, J48 classifier was to be the classify for classification.

First test -> Correctly Classified Instances 0%  (Amazing!!!) -> Change data sets by running Sequitur each 60 characters (one instance was composed of bag 0f words of  60 characters per a class) -> Second test Correctly Classified Instances 41.1765% (Wow!!!).

During the week, I have drawn a proposed research framework (via Visio) in order that my points of view will be clear and I will also know how narrow/broad/scope of my research.

Then, the research framework was exactly clear and obviously direction.  These things were happened after meeting today (08-09-09).   Happy!!!

According to Nevill-Manning & Witten (1997) stated that “SEQUITUR is an algorithm that infers a hierarchical structure from a sequence of discrete symbols by replacing repeated phrases with a grammatical rule that generates the phrase, and continuing this process recursively.“

 

Ref. Nevill-Manning & Witten (1997). Identifying Hierarchical Structure in Sequences: A linear-time algorithm. Journal of Artificial Intelligence Research, 7, pp. 67-82.

Dear Georgetown Cuddler,

Or would it be Dear THE Georgetown Cuddler, or quite possibly, Cuddlers.  You have been quite the busy little guy over this past year.  I say this past year because that’s when you entered my consciousness, but it’s possible you’ve been around cuddling forever. Maybe you’re kind of like the Dread Pirate Roberts of creepsters, passing down  the mantle once you get old and enfeebled  to a younger, much more spry creepster, more capable of sneaking into unlocked Georgetown apartments and row houses.  For the uninitiated, lemme ’splain.  Or rather, let me let the good folks over at Georgetown University ’splain:

Dear Members of the Georgetown University Community,

Please be advised that on Sunday, August 30, 2009 at approximately 6:30 a.m., an unidentified male entered a student’s residence in the 1200 Block of 33rd Street, NW, and lay down on the couch with her. The complainant was startled awake, at which time she told the suspect to leave, which he subsequently did.

And this is by no means the first occurrence.  At least six reported cases in the last calendar year alone.  So that’s this guys deal.  He doesn’t steal.  He doesn’t assault.  He doesn’t watch TV like John Travolta  in that one scene from Get Shorty.  He doesn’t eat your last Babybell soft cheese or drink the last beer.  He cuddles.

Some of you who know me may be familiar with my love of Arlo Guthrie (Hold on, this will eventually turn in to a sequitur).  On his self titled album, he has a track known as the “Pause of Mr. Clause.” The song itself is a two minute nothing, tucked away on the last track of the B side, a fun christmas song about how Santa is a pacifist (long hair and beard) and a communist (red suit) and a dope fiend (sneaks into your house at night, smokes something from a pipe). What makes the track amazing is the 7 minute discussion of the FBI that precedes the song, during which Arlo talks about (among other things) how everyone says, “Well, my life might stink, but look at that guy, I’m better off than him,” and then proceeds to talk about the last guy.  The guy with no one to point to.  (See it, now?)  I think we may have found the last guy.

So, Georgetown Cuddler, or Cuddler for short, what’s up?  Obviously, you’re lonely, we get that.  A bit eccentric.  But, why do you need to literally steal human contact?  I guess my question is how did you get this way.  At what point does one decide that they need a cuddle so bad that they’re willing to commit a felony (or is it a misdemeanor?) for the touch of another human being? Why is it so hard for you to get close to people? (Chronic halitosis, schizophrenia … libertarian?)

I know I’m asking you a lot of questions, and I don’t want to seem accusatory or judgmental.  As a fairly tactile dude, I know how great a good cuddle can feel.  I get that.  I’m with you.  I know.  But, I think I’ve got a solution for you.  A way for you to connect with other people, and maybe start focusing your attention on women who might be a little more receptive.

Cuddles, it’s time you got on craigslist or THE craigslist, depending on your age.  I know it might be scary to put yourself out there on the big wide internets, especially for the first time. So, being the great guy that I am, I’ve decided to help you out.  All you’ve gotta do is copy the following ad, put it on craigslist and  - Voila! – legitimate cuddling.  Don’t change a word though, because I’ve done my best to make it as non-threatening as possible, something I think you may have had some trouble with in the past.  OK, here goes.

Hey Baby, Let’s Cuddle

Do you ever find yourself pining for someone to hold you? Someone to listen to you breathe and stand over your bed and stare at you while you sleep? Well, look no more.  Something-something white male seeks twenty-something female for cuddling.   Prefer little spoon, but can be big spoon if it’s a deal-breaker.  I don’t really want to do anything else.  I mean that.  No dates, no movies, no talking, just some good old fashioned, no strings cuddling.  I operate mostly in the Georgetown area.  All interested women should leave their doors unlocked at night, that way I’ll know to just go ahead and let myself in.  NO DUDES!

• cats are OK – purrr
• dogs are OK – wooof
• Location: Georgetown
• it’s NOT ok to contact this poster with services or other commercial interests

So there you have it, Cuddles.  And trust me, that would by no means be anywhere near the weirdest thing ever put on craigslist.  Not by a long shot.

Keep on Truckin’,

- Timmy

P.S – I saw some dudes in Free Hugs t-shirts hanging outside of some bars on M St.  If you need a quick fix, they may be a good place to start.  Just make sure you tell ‘em “No Homo” right before you move in for the kill.

The weekly meeting has been rescheduled from Tuesday to Friday at 1.00 p.m. and this will extend my studying for this week as well.

During this week, I have learned about StringToWordVector filter in Preprocess, Bag of Words and String Attribute in ARFF format. I also learned from supervisor about using various WEKA functions e.g. Classify, Search Method in Select attributes i.e. ExhaustiveSearch, GreedyStepwise, and Attribute Evaluator. Preparing suitably training sets is to create appropriately models. Afterthat, test sets will be dealt with these models.

Huge Attribute Set Small Attribute Set
Gene Sequence —->  All Substrings  —–> Attribute Selection ——> Classify
|                             (candidate attributes)    (WEKA/TF*IDF)               |
|——————————–SEQUITUR ———————————————————|

To do lists for next week: prepare new training/test sets (use Sequitur), learn the various functions of  Select attributes (in Weka).

Note! ExhaustiveSearch will take long time evaluating on training data.

Quit trying to change the subject, VPL. You’re a tool belt…

(photo courtesy of Josh from dbwithbt)

I have reported what I did last week.

1. Counting the set of 10 thru all of samples was prepared for data collection. I have dealt these data with awk one command-line (e.g. $awk -Fa ‘{print NF-1} ‘ lmg.data) and kept those number of characters (A,C,G,T) in the excel format. In order that I was able to calculate the percentage of these characters and prepared the data entry in ARFF format to run with WEKA later.
2. Choosing which algorithms are suitable for these data was the main point to read and study to understand various algorithms.

My supervisor has suggested me a new tool ‘Sequitur‘ (http://www.sequitur.info) and explained the Naive Bayes classifier.

To Do Lists:
1. Prepare data in sequitur format and understand these output.
2. Read more Naive Bayes classifier and Decision Tree.

Obama: The Great Non Sequitur · Link. What

 MUDGE’s Musings  

There’s most read, and then there’s favorite. This is a post which yr (justifiably) humble svt is, regrettably, but not regretfully, not at all humble about.

Blast from the Past!

A post we really, really loved to write, and read, and re-read…

From our early days, originally posted August 27, 2007.

mm119: Creating the sequitur

Had this thought yesterday.

Any of you regular reader of this nanocorner of the blogosphere are aware that MUDGE often is slightly link-crazy.

I believe I learned this style best from one of my most regular reads, Slate.com, and good teachers they’ve been.

What linking does for yours truly, and here comes that flash of insight –drum-roll please — linking sequiturizes.

We all think we know what a non sequitur is. My new toy WordWeb says that it’s “A reply that has no relevance to what preceded it.”

So, when MUDGE pops out a comment, like this one from a recent posting…

An ambitious plan, to be executed of course by the lowest bidder.

… you don’t have to be in the dark about what I might mean by that.

Because I’ve provided a handy hyperlink to let you know exactly what I had in mind.

Wikipedia’s article goes a bit into the history of the hyperlink:

The term “hyperlink” was coined in 1965 (or possibly 1964) by Ted Nelson at the start of Project Xanadu. Nelson had been inspired by “As We May Think,” a popular essay by Vannevar Bush. In the essay, Bush described a microfilm-based machine (the Memex) in which one could link any two pages of information into a “trail” of related information, and then scroll back and forth among pages in a trail as if they were on a single microfilm reel. The closest contemporary analogy would be to build a list of bookmarks to topically related Web pages and then allow the user to scroll forward and backward through the list.

Ted Nelson was a genius — I remember reading him in the fondly remembered Byte Magazine (I was a near-charter subscriber, of course), and thinking “this man is a way out futurist.”

So, the non sequiturs will keep on coming, gang, sequiturized (must be the process of negating a non sequitur, right), by Ted Nelson’s (and of course Tim Berners-Lee’s) marvelous hyperlinks.

New motto for this nanocorner:

It’s it for now. Thanks,

–MUDGE

MUDGE’S Musings

Had this thought yesterday.

Any of you regular reader of this nanocorner of the blogosphere are aware that MUDGE often is slightly link-crazy.

I believe I learned this style best from one of my most regular reads, Slate.com, and good teachers they’ve been.

What linking does for yours truly, and here comes that flash of insight –drum-roll please — linking sequiturizes.

We all think we know what a non sequitur is. My new toy WordWeb says that it’s “A reply that has no relevance to what preceded it.”

So, when MUDGE pops out a comment, like this one from a recent posting…

An ambitious plan, to be executed of course by the lowest bidder.

… you don’t have to be in the dark about what I might mean by that.

Because I’ve provided a handy hyperlink to let you know exactly what I had in mind.

Wikipedia’s article goes a bit into the history of the hyperlink:

The term “hyperlink” was coined in 1965 (or possibly 1964) by Ted Nelson at the start of Project Xanadu. Nelson had been inspired by “As We May Think,” a popular essay by Vannevar Bush. In the essay, Bush described a microfilm-based machine (the Memex) in which one could link any two pages of information into a “trail” of related information, and then scroll back and forth among pages in a trail as if they were on a single microfilm reel. The closest contemporary analogy would be to build a list of bookmarks to topically related Web pages and then allow the user to scroll forward and backward through the list.

Ted Nelson was a genius — I remember reading him in the fondly remembered Byte Magazine (I was a near-charter subscriber, of course), and thinking “this man is a way out futurist.”

So, the non sequiturs will keep on coming, gang, sequiturized (must be the process of negating a non sequitur, right), by Ted Nelson’s (and of course Tim Berners-Lee’s) marvelous hyperlinks.

New motto for this nanocorner:

It’s it for now. Thanks,

–MUDGE

During week5-week6, I was summarizing my project overview and graphically represent my point of view to  supervisor via freemind. After a long discussion in meeting day, the assignment for this week is a comparison of Sequitur and Keyphrase Extraction Algorithm (KEA) working function.

From Wikipedia,  Sequitur (or Nevill-Manning algorithm) is a recursive algorithm developed by Craig Nevill-Manning and Ian H. Witten in 1997^[1] that infers a hierarchical structure (context-free grammar) from a sequence of discrete symbols. The algorithm operates in linear space and time. It can be used in data compression software applications.

“Sequitur is a method for inferring compositional hierachies from strings. It detects repetition and factors it out of the string by forming rules in a grammar. The rules can be composed of non-terminals, giving rise to a hierarchy. It is useful for recognizing lexical structure in strings, and excels at very long sequences“  (This text is copied from http://www.sequitur.info)

“Keyphrase Extraction Algorithm (KEA) is an algorithm for extracting keyphrases from text documents. It can be either used for free indexing or for indexing with a controlled vocabulary.” (This is copied from http://www.nzdl.org/Kea)