Frequent Pattern Trees Research paper implementation

[Ahmedhussain]

Hi,

I am working on a Job portal which caters its jobseekers interests. Now I have a research paper that I have to implement on it. and its based on Frequent Pattern Trees.

Now I tried google checked few websites, but didnt understand how it worked and what could be the implementation of this tree and the rest of the algorithm.

Can any one provide help on this. I have upload an attachment of a research paper so that in case if any one can help in this regard, that individual can have the reference of what I am talking about ... :)

Thanks in advance

Regards,

Syed Ahmed Hussain

[Airslash]

simply put, and I mean really simple, the paper states that if you wish to provide an online portal or community platform you have to keep several things in mind:

- people don't want supervision, they want the illusion to work uncontrolled
- people will have a standard set of expectations
- people want freedom
- provide a general platform with minimal support, but allow custom content and customizations
- do not force things on your customers.

The paper is just a mathematical approach with statistics to explain the above points. However, any reasonable web programmer should understand the quircks of the Internet and it's users.

[Ahmedhussain]

Hi Airslash,

Thanks for the reply. I totally agree to your points,
but how will I be able to implement this mathematical approach?
And what will be the implementation of Frequent Pattern tree? or is there any other alternative to handle it.. i.e. with any other datastructure?

Thanks,

Regards,

Syed Ahmed Hussain

[Airslash]

your best approach would be by constructing the minimalistic platform.
E.g a simple webbased platform that supports the basic functionality.

ofcourse you would have to design for yourself what this basic functionality would be. All I can give as a tip, is that the basic functionality on a portal; is that what all members would have in common.
Implementing that would cover the mathematical part of the similarity & Union.

Once that is finished, you'd have to focus on the needs and wishes of the majority of the community. Find what they need/want and what applies to at least 65% of the total users. That covers the majoraty part of the math.

Last, but not least, you can create progeammable API & interfaces that allow customization by the users. Examples are custom themes, layouts and styles.
That would cover the individuality of the tree.

=============== =============== =============== =============== ===
As for implementation, it heavily depends on the platform and programming language you intend to use and functionality of your underlying platform.
If you intend to work with ADO.NET and ASP.NET pages then the DAO & fatory patterns would definitly suit you.

If you on the other hand tend to lean towards Ruby on Rails forexample, I think STI 'Single Table Inheritance' would be a nice way to go.

I think there are alot of different ways to implement this; it's been a while since I done some webdeveloping.

[jhumelsine]

I looked over the paper. Wow. The word "obfuscate" comes to mind.

I don't know much about FP Trees, but I did find a link that might help: http://www.csc.liv.ac.uk/~frans/KDD/.../fpGrowth.html

I think it might be close to what you're looking for. It includes links to a Java implemenation, which should help, if what the link is talking about is anywhere close to what you're working upon.

If not, make sure to do a Google image search in addition to a web search. Sometimes I can find what I'm looking for faster by an image than a description. I found this link via an image search.

[Ahmedhussain]

your best approach would be by constructing the minimalistic platform.
E.g a simple webbased platform that supports the basic functionality.

ofcourse you would have to design for yourself what this basic functionality would be. All I can give as a tip, is that the basic functionality on a portal; is that what all members would have in common.
Implementing that would cover the mathematical part of the similarity & Union.

Once that is finished, you'd have to focus on the needs and wishes of the majority of the community. Find what they need/want and what applies to at least 65% of the total users. That covers the majoraty part of the math.

Last, but not least, you can create progeammable API & interfaces that allow customization by the users. Examples are custom themes, layouts and styles.
That would cover the individuality of the tree.

=============== =============== =============== =============== ===
As for implementation, it heavily depends on the platform and programming language you intend to use and functionality of your underlying platform.
If you intend to work with ADO.NET and ASP.NET pages then the DAO & fatory patterns would definitly suit you.

If you on the other hand tend to lean towards Ruby on Rails forexample, I think STI 'Single Table Inheritance' would be a nice way to go.

I think there are alot of different ways to implement this; it's been a while since I done some webdeveloping.

Well thankyou very much Airslash...
I am using ASP.NET (C#)
please correct me if I am wrong.

You mean to say that I should create a dataset and on that dataset I directly implement the algorithm?

[Ahmedhussain]

I looked over the paper. Wow. The word "obfuscate" comes to mind.

I don't know much about FP Trees, but I did find a link that might help: http://www.csc.liv.ac.uk/~frans/KDD/.../fpGrowth.html

I think it might be close to what you're looking for. It includes links to a Java implemenation, which should help, if what the link is talking about is anywhere close to what you're working upon.

If not, make sure to do a Google image search in addition to a web search. Sometimes I can find what I'm looking for faster by an image than a description. I found this link via an image search.

hi jhumelsine,

Thanks you very much for providing me this..

It is really helpfull. But Im not good in java... C# programmer :(