The Data Mining Forum                             open-source data mining software data science journal data mining conferences high utility mining book
This forum is about data mining, data science and big data: algorithms, source code, datasets, implementations, optimizations, etc. You are welcome to post call for papers, data mining job ads, link to source code of data mining algorithms or anything else related to data mining. The forum is hosted by P. Fournier-Viger. No registration is required to use this forum!.  
source code of EFIM for high utility itemset mining
Posted by: Hasim
Date: October 22, 2015 09:26AM

Hello: I try to understand the code of EFIM for high utility mining and why it is faster than UPGrowth, FHM, and HUI-Miner. Could you explain to me?



Edited 1 time(s). Last edit at 12/27/2015 04:42AM by webmasterphilfv.

Options: ReplyQuote
Re: source code of EFIM for utility itemset mining
Date: October 23, 2015 05:47PM

Hi,

I will not explain the algorithm into details because it would be too long. But the major features of EFIM are:
- EFIM is a one phase algorithm for high-utility itemset mining. Thus it does not generate candidates unlike UPGrowth, TwoPhase and such algorithms
- Unlike algorithms such as HUI-Miner and D2HUP, EFIM will calculate the remaining utility (called sub-tree utility) at the parent node rather that at child nodes in the depth-first search, thus it can prune more nodes in the search space. Moreover, the upper-bound in EFIM is better as locally unpromising items are removed from the sub-tree utility. In FHM and HUI-Miner, locally unpromising items cannot be removed from the remaining utility since these algorithms use a vertical database. Thus, sub-tree utility is a tighter upper bound than the remaining utility used in HUI-Miner and D2HUP and generally also better than algorithms using the TWU such as UPGrowth+, UPGrowth, IHUP and Two-Phase.
- EFIM utilizes the idea of transaction merging which is hard or very difficult to implement in HUI-Miner and D2HUP since they are respectively vertical algorithm and hyperlink-based algorithms. Transaction merging greatly improve the runtime.
- A key design decision in EFIM is to perform all operations for processing an itemset in the search space in linear time, while utility-lists in HUI-Miner can be constructed in O(n^3) time in some case.
- ...

By all these strategies and optimizations, as you can see in the experiments, EFIM can be up to 1000 times faster than the state-of-the-art HUI-Miner, FHM, UPgrowth and D2HUP algorithms. Also, another key feature of EFIM is that it is very memory efficient. It can use up to eight times less memory than the previous best algorithms such as HUI-Miner and FHM. This is due to the use of simple data structures.



So this is the main idea. You can get more information in the paper describing EFIM published at MICAI 2015:

Zida, S., Fournier-Viger, P., Lin, J. C.-W., Wu, C.-W., Tseng, V.S. (2015). EFIM: A Highly Efficient Algorithm for High-Utility Itemset Mining. Proceedings of the 14th Mexican Intern. Conference on Artificial Intelligence (MICAI 2015), Springer LNAI, to appear.

Also, if you send me an e-mail, I can send you an extended version of the paper with more details about the optimizations used in EFIM, and also more experiments.

Hope that this answer your question.

Philippe



Edited 9 time(s). Last edit at 12/14/2015 09:00PM by webmasterphilfv.

Options: ReplyQuote
Re: source code of EFIM for utility itemset mining
Posted by: venkatachalam
Date: December 14, 2015 08:52PM

Hi,
I am an research scholar doing my research in high utility item set mining, if you could send me the details and optimization used in EFIM: A Highly Efficient Algorithm for High-Utility Itemset Mining it will be useful to me. please send it to my email kavitha.psk@gmail.com

Options: ReplyQuote
Re: source code of EFIM for utility itemset mining
Date: December 14, 2015 08:57PM

Hi,

You can download the conference paper about EFIM here:

http://www.philippe-fournier-viger.com/MICAI2015_EFIM_High_Utility_Itemset_Mining.pdf

and you can get the Java source code of EFIM here (as part of the SPMF data mining library):

http://www.philippe-fournier-viger.com/spmf/

and you can get the datasets here:

http://www.philippe-fournier-viger.com/spmf/index.php?link=datasets.php

Besides, if you have some specific questions about EFIM, you can always ask here or send me an e-mail.



Edited 2 time(s). Last edit at 12/14/2015 08:58PM by webmasterphilfv.

Options: ReplyQuote
Source code of EFIM for utility itemset mining with FHM+ length constraint
Posted by: Shamna. K.K
Date: September 21, 2017 02:20AM

Hi,
I am doing M.tech.My project is in high utility item set mining,Please help me to alter EFIM algorithm with length constarint(used in FHM+ algorithm).Please send it to my email id :shamnahakheemvp@gmail.com

Options: ReplyQuote
Re: source code of EFIM for utility itemset mining
Posted by: Samy
Date: August 01, 2016 05:32AM

Hi sir, can you please the share the source code for d2HUP algorithm or EFIM....

Options: ReplyQuote
Re: source code of EFIM for utility itemset mining
Date: August 14, 2016 04:48AM

Hi,

You can get the Java source code of EFIM and d2hup in the SPMF data mining library. See the link on top of this page. EFIM was shown to be faster than d2hup, upfront and FHM.

Philippe

Options: ReplyQuote
Re: source code of EFIM for high utility itemset mining
Posted by: venkatesh
Date: December 26, 2016 04:56PM

PLZ SEND ME PROJECT

Options: ReplyQuote
Re: source code of EFIM for high utility itemset mining
Date: December 26, 2016 06:26PM

You can get the source code from the download page of SPMF:

http://www.philippe-fournier-viger.com/spmf/

Options: ReplyQuote
Re: source code of EFIM for high utility itemset mining
Posted by: Moham
Date: August 25, 2017 09:58PM

Please give me the source code of EFIM for my research

Options: ReplyQuote


Your Name: 
Your Email: 
Subject: 
Spam prevention:
Please, enter the code that you see below in the input field. This is for blocking bots that try to post this form automatically.
 **        ********   ********   ******         ** 
 **        **     **  **    **  **    **        ** 
 **        **     **      **    **              ** 
 **        ********      **     **              ** 
 **        **           **      **        **    ** 
 **        **           **      **    **  **    ** 
 ********  **           **       ******    ******  
This forum is powered by Phorum and provided by P. Fournier-Viger (© 2012).
Terms of use.