The Data Mining Forum                             open-source data mining software data mining conferences Data Science for Social and Behavioral Analytics DSSBA 2022 data science journal
IMPORTANT: This is the old Data Mining forum.
I keep it online so that you can read the old messages.

Please post your new messages in the new forum: https://forum2.philippe-fournier-viger.com/index.php
 
fast maximal frequents itemset algorithm
Posted by: djanahana
Date: June 21, 2013 02:44PM

Hi,

please i want to know the list of efficient and fast maximal frequents itemset extraction algorithm.

i find GenMax as an efficient one,is it the only one.

another question, i read some paper where the authors developpe (make the source code of some data mining algorithm "FP max as an example " and they publish a paper

is it a search axe if i developpe an algorithm.


thanks for reading my message

Options: ReplyQuote
Re: fast maximal frequents itemset algorithm
Date: June 21, 2013 05:25PM

Hi Djanahana,

For the first question, I think that the main algorithms for maximal frequent itemset mining are:

MAFIAX, GENMAX, FPMAX, LCM, MAXMINER, SMARTMINER, ...

But maybe that I have missed some. I did not do an extensive search on this topic. You can may be find more on Google Scholar.

For the second question, implementing an existing algorithm is generally not enough to publish a paper unless you improve with some new significant optimizations or data structures and show that it performs better in some conditions. Another possibility is to modify the algorithm so that it does something slightly different. For example, FPMAX is an algorithm for mining maximal itemset. Could we modify it to mine a slightly different kind of itemsets such as maximal fuzzy itemsets? maximal uncertain itemset? etc. That could be interesting. What I mean is that you need to improve the algorithm or show that you add something new.

Hope this helps,

By the way, thank you for updating the conference list!

Philippe

Options: ReplyQuote
Re: fast maximal frequents itemset algorithm
Posted by: djanahana
Date: June 22, 2013 01:58PM

Hi ,

Thanks for the list of algorithms, for the paper i was speaking about as an example this paper
http://dl.acm.org/citation.cfm?id=1802277

Sir could you please tell me where can i find a java source code of any of these algorithm (if there is). I see that you have developped CHARM-MFI but as you said it isn't an efficient one.


Thanks in advance,


Djanahana

Options: ReplyQuote
Re: fast maximal frequents itemset algorithm
Date: June 22, 2013 03:17PM

Hi,

I have spent 20 minutes searching on Google but cannot find it.

Here is two ideas:
- You could contact the authors of the paper about the Java implementation to ask them if they can send you the code.
- Alternatively, I have found this website that seems to have the code: http://en.pudn.com/downloads432/sourcecode/java/detail1827711_en.html But to get the code they ask either to (1) upload five files or (2) to pay.

Otherwise, another solution is to implement it by yourself. But it may be difficult.

Best,

Philippe

Options: ReplyQuote
Re: fast maximal frequents itemset algorithm
Posted by: djanahana
Date: June 23, 2013 03:05AM

Hi,

I am sorry for waisting your time i have already look for the source code of FP-max, and i couldn't find it.


I was looking for a java source code of any other algorithm for maximal frquent itemset algorithm generation



Many thanks for your time


thanks,


Best regards;

Options: ReplyQuote
Re: fast maximal frequents itemset algorithm
Date: June 23, 2013 04:41AM

Hi again,

The only thing that I can find is this:

code.google.coém/p/fci-dr/

It offers an implementation of LCM but I think that it only discovers closed itemsets. I think that the author did not implement the part about maximal itemsets in LCM. I have not tested it. But maybe that it would be possible to modify it to find maximal itemsets. I don't know.

Best,

Philippe

Options: ReplyQuote
Re: fast maximal frequents itemset algorithm
Date: June 23, 2013 04:57AM

Update:

I have make a quick test to see how good this LCM implementation is and it does not seems very fast. I have compared LCM and my implementation of CHARM and DCI-Closed for mining closed itemsets on the Chess dataset with minsup = 85 % and my implementations of CHARM and DCI_Closed finish in less than 1 second, while the LCM code takes about 10 seconds... Moreover, the LCM code does not generate the same amount of results (i don't know why : bug? maximal itemsets? i don't know)

C:\Users\ph\Desktop\SPMF>java -jar spmf.jar run Charm_bitset chess.txt test.txt
85%
============= CHARM - STATS =============
Transactions count from database : 3196
Frequent closed itemsets count : 1885
Total time ~ 800 ms
===================================================

C:\Users\ph\Desktop\SPMF>java -jar lcm-java.jar chess.txt 85
=====================================
Total count: 1672
=====================================

Options: ReplyQuote
Re: fast maximal frequents itemset algorithm
Posted by: djanahana
Date: June 23, 2013 10:03AM

Hi ,


I truly appreciate your effort and kindness to help me.


it seems very difficult to get a source code of those algorithms.confused smiley



thanks once again


Best regards;

Options: ReplyQuote
Re: fast maximal frequents itemset algorithm
Date: June 23, 2013 10:19AM

I agree... It seems also that there are not a lot of recent papers on this topic. But it is an interesting topic.

Philippe

Options: ReplyQuote
Re: fast maximal frequents itemset algorithm
Posted by: surendra
Date: May 07, 2014 11:01PM

HELLO EVERY ONE,CAN YOU TELL ME HOW TO USE ANT COLONY ALGORITHM TO FINDING THE FREQUENT ITEMSET IN JAVA

Options: ReplyQuote


This forum is powered by Phorum and provided by P. Fournier-Viger (© 2012).
Terms of use.