Word Clustring project basis

Question

1.00/5 (3 votes)

See more:

, +

I'm thinking about my graduation project to be word clustring project but i don't know what knowledge I should have before starting
I'm taking pattern recognition & classification course at college now
any help?
thanks in advanced

Posted 28-Oct-14 9:01am

Mujeeba Haj Najeeb

Add a Solution

2 solutions

Add a Solution

Add your solution here

Treat my content as plain text, not as HTML

Preview 0

…

Existing Members

Sign in to your account

...or Join us

Download, Vote, Comment, Publish.

Your Email
Password
Forgot your password?

Your Email
This email is in use. Do you need your password?
Optional Password

I have read and agree to the Terms of Service and Privacy Policy
Please subscribe me to the CodeProject newsletters

When answering a question please:

Read the question carefully.
Understand that English isn't everyone's first language so be lenient of bad spelling and grammar.
If a question is poorly phrased then either ask for clarification, ignore it, or edit the question and fix the problem. Insults are not welcome.
Don't tell someone to read the manual. Chances are they have and don't get it. Provide an answer or move on to the next question.

Let's work to help developers, not make them feel stupid.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

Sergey Alexandrovich Kryukov · Accepted Answer · 2014-10-28T09:22:00

Solution 1

Why so many people ask about "starting"? There are too many ways to start, and many of those ways are good enough. This is not so critical which one you would choose.
For example, you can start with learning relevant part of mathematics: https://en.wikipedia.org/wiki/Cluster_analysis[^].

First, very elementary mathematical thing is this: to do clustering, you should have some space of elements of your set you want to cluster, and that space should have a correctly defined norm, to make a normal space: https://en.wikipedia.org/wiki/Normed_vector_space[^].

In simple words, you should have (and supply to the calculation algorithm) the function which calculates the distance between two objects (which you called words, in your case). Such functions could be different; but this is not a fully arbitrary function; it should behave to meet the requirements posed by the axioms of the metric space. Only then you can apply the cluster analysis to your set. The article referenced above is really just the staring point.

—SA

Posted 28-Oct-14 9:22am

Sergey Alexandrovich Kryukov

Comments

Mujeeba Haj Najeeb 2-Nov-14 9:40am

I need to know the concept of word/text clustring
I haven't realise it well
e.g: in OCR we process a text to make letters extraction
my question is: what do we do in word/text clustring?

Sergey Alexandrovich Kryukov 2-Nov-14 15:59pm

As I say, you define what is the distance between words, called "norm" for the space. No matter what is it, but it can meet the axioms. Say if A > B and B > C, then should be A > C (transitive rule), and so on. Just read on the topic. And then follow the general theory of clustering, which, from this point is abstracted from the nature of elements you want to cluster, as soon ans the distance (norm) is correctly defined and works.
—SA

Mujeeba Haj Najeeb 5-Nov-14 14:36pm

I'll explain my question in another way
which applications need to use word/text clustring?
I need ideas to start

Sergey Alexandrovich Kryukov 5-Nov-14 16:12pm

Do you want to implement clustering in some application of other software product, or do you need to understand why going for that at all?
If by "application" you mean using some available application, you are at the wrong forum. We usually advise on how to create software.
—SA

Mujeeba Haj Najeeb 6-Nov-14 14:56pm

first of all I need to understand why going for that at all?
and I also need to know some applications in real life to realize it more specifically
I need to know where am going, I need a general perspective to make a decision to enter in such field or not

Sergey Alexandrovich Kryukov 6-Nov-14 15:26pm

This is a very broad subject; I answered in a separate post, Solution 2.
—SA

Mujeeba Haj Najeeb 7-Nov-14 2:10am

OKay

Sergey Alexandrovich Kryukov · Accepted Answer · 2014-11-06T09:25:00

Mujeeba Haj Najeeb asked:

first of all I need to understand why going for that at all?
and I also need to know some applications in real life to realize it more specifically
I need to know where am going, I need a general perspective to make a decision to enter in such field or not.

Thank you for your clarification; fair enough.

It's hard to cover any comprehensive applications of this kind of analysis, but, generally, it is used in some fields of linguistics, in particular, natural language processing:
http://en.wikipedia.org/wiki/Natural_language_processing[^].

It can be considered as one of the many parts of computational linguistics:
http://en.wikipedia.org/wiki/Computational_linguistics[^].

See also:
http://en.wikipedia.org/wiki/Word-sense_induction[^],
http://en.wikipedia.org/wiki/Ambiguity[^].

Note that in the examples mentioned above, the norm (distance) itself (I discussed the role of norm in Solution 1) is extremely complex: it should reflect semantic similarity a very complex notion which is itself very hard to formalize. The norm values for some word set (thesaurus) can come from extensive statistical analysis, expert systems, and the like. It has nothing to do with the string comparison algorithms implemented in most libraries.

See also: http://en.wikipedia.org/wiki/Word-sense_disambiguation[^].

Maybe there are different applications which I never heard of and could only speculate about. For example, one of my friends specialized in computational linguistic and defended his dissertation on such thing as inferring individual characteristics of a writer based exclusively on statistics of the words found in the text samples.

I must say that computational linguistics is a developing branch of science which is not yet really close to, say, serious commercial use. I feel that the major bread-through works lie in future, which might look attractive to the newcomers. I believe every serious work in this field is on the cutting edge of both linguistics and applied mathematics (and maybe even "fundamental" mathematics. If you want to go for it (and it's good that you asked this question), you need to have solid mathematical background and seriously go into linguistic, which is really hard to do. I don't think that being "just a programmer", a part of technical stuff, can be practical or reasonable. Getting into real science is the only thing which makes sense, but this is not for everyone, so try to be realistic. I hope you won't consider my words as discouragement. I would be more than happy to know that you take this route and are successful.

—SA