assignment 2
Posted by Anonymous User
| Announcements | Last Post | |
|---|---|---|
 |
SoC Curricula | 09/30/2017 01:08PM |
|
Demarcation or scoping of examinations and assessment | 02/13/2017 07:59AM |
|
School of Computing Short Learning Programmes | 11/24/2014 08:37AM |
|
Unisa contact information | 07/28/2011 01:28PM |
|
Re: assignment 2 May 23, 2007 12:44PM |
thanks! 
|
Re: assignment 2 May 25, 2007 01:00PM |
did you conrtact her via the email address on the phorums or the one we are supposed to email our assignments to?
|
May 30, 2007 12:05PM |
|
Re: assignment 2 May 31, 2007 03:44PM |
i love recursion... but i don't get what mangling a bunch of letters into various combinations has to do with language processing... but maybe i expect too much.
|
June 04, 2007 08:04AM |
I finally got around to reading chapter 3 and this is my take on things.Quote
i love recursion... but i don't get what mangling a bunch of letters into various combinations has to do with language processing... but maybe i expect too much.
We are not meant to use a dictionary that has every single possible word in it (as this would be far too easy, large and as you have said have nothing to do with language processing)
Instead we are meant to use a stem dictionary and construct a FST and use this to validate the words.
I could be wrong but I think this is the case, anyone else care to comment?
--
"Knowledge has much better uses than self-pity and superiority"
|
June 04, 2007 08:46AM |
|
Re: assignment 2 June 04, 2007 12:00PM |
malcolm Wrote:
-------------------------------------------------------
> I finally got around to reading chapter 3 and this
> is my take on things.
> We are not meant to use a dictionary that has
> every single possible word in it (as this would be
> far too easy, large and as you have said have
> nothing to do with language processing)
> Instead we are meant to use a stem dictionary and
> construct a FST and use this to validate the
> words.
>
> I could be wrong but I think this is the case,
> anyone else care to comment?
i don't think so... chapter 3 seems to say that a FST for english is massively complex... also, i'm not sure of any decent strategy to generate anagrams using all the letters in the original word and how a FST would fit into that.
-------------------------------------------------------
> I finally got around to reading chapter 3 and this
> is my take on things.
> We are not meant to use a dictionary that has
> every single possible word in it (as this would be
> far too easy, large and as you have said have
> nothing to do with language processing)
> Instead we are meant to use a stem dictionary and
> construct a FST and use this to validate the
> words.
>
> I could be wrong but I think this is the case,
> anyone else care to comment?
i don't think so... chapter 3 seems to say that a FST for english is massively complex... also, i'm not sure of any decent strategy to generate anagrams using all the letters in the original word and how a FST would fit into that.
|
June 04, 2007 12:21PM |
You would still generate all possible combinations, but in order to check if they are valid words you would run them against the FST instead of against a giant dictionary.
It would be both A) faster and
take up far less space on disk.
As for it being complex, this is an honours subject right? ;/
--
"Knowledge has much better uses than self-pity and superiority"
It would be both A) faster and
take up far less space on disk.
As for it being complex, this is an honours subject right? ;/
--
"Knowledge has much better uses than self-pity and superiority"
|
Re: assignment 2 June 04, 2007 03:37PM |
malcolm Wrote:
-------------------------------------------------------
> You would still generate all possible
> combinations, but in order to check if they are
> valid words you would run them against the FST
> instead of against a giant dictionary.
> It would be both A) faster and take up far less
> space on disk.
>
> As for it being complex, this is an honours
> subject right? ;/
hrm.... you might be right - but based on how i perceived the question that is not what is required.
i don't know if the FST would be faster tho... a well index in memory dictionary (or just a collection like a java set) would take up a lot more memory but would be lightening fast on query.
-------------------------------------------------------
> You would still generate all possible
> combinations, but in order to check if they are
> valid words you would run them against the FST
> instead of against a giant dictionary.
> It would be both A) faster and
take up far less
> space on disk.
>
> As for it being complex, this is an honours
> subject right? ;/
hrm.... you might be right - but based on how i perceived the question that is not what is required.
i don't know if the FST would be faster tho... a well index in memory dictionary (or just a collection like a java set) would take up a lot more memory but would be lightening fast on query.
|
June 04, 2007 10:24PM |
Or which operating system it is for
--
"Knowledge has much better uses than self-pity and superiority"
--
"Knowledge has much better uses than self-pity and superiority"
|
June 05, 2007 10:18AM |
|
June 05, 2007 10:25AM |
The assignment lists the "on-line" Dictionary as one of the 5 files we must submit, which I guess means it should infact be a file and not a real online dictionary? 
--
"Knowledge has much better uses than self-pity and superiority"
--
"Knowledge has much better uses than self-pity and superiority"
|
June 05, 2007 11:15AM |
|
Re: assignment 2 June 05, 2007 11:24AM |
and start worrying about assignment 3 
Sorry, only registered users may post in this forum.