The Data Mining Forum                             open-source data mining software data science journal data mining conferences high utility mining workshop
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!.  
generate PMML output
Posted by: Diep Dao
Date: September 11, 2018 08:45AM

Hi,

I have been using SPMF algorithms for my latest research and I really appreciate your work on this. I want to ask if you would suggest a good way to get the mined rule in PMML format?

Many thanks,
Diep Dao

Options: ReplyQuote
Re: generate PMML output
Date: September 11, 2018 04:03PM

Hi,

Thanks for using SPMF. I do not know what is the PMML format. But if you are comfortable with Java, you could modify the code for writing the rules to the file. This should not be hard. In SPMF, each algorithm is in a separated package. So you could first find the code of the algorithm that you want to modify and then change the code.

But what is PMML? Can you give me a link to a website or document describing it? I can have a look too.

Best,

Philippe

Options: ReplyQuote
Re: generate PMML output
Posted by: Dang Nguyen
Date: September 11, 2018 07:56PM

@Philippe: PMML should be Predictive Model Markup Language. You can refer to this link: https://en.wikipedia.org/wiki/Predictive_Model_Markup_Language

@Diep: If you need, I share my code (in C#) which writes rules to the PMML format.

Options: ReplyQuote
Re: generate PMML output
Date: September 13, 2018 08:29AM

Thanks Dang,

I see. An XML format. I don't like too much XML-based format personally. It wastes a lot of space with all these tags. Already the text-based format of SPMF takes a lot of space because it is a text file. This format would maybe make the output files 10 times or more larger. Just my opinion. But I understand that it can be useful for interoperability with other software.

Is it widely used this PMML format? Do you think it would be useful to useful for many people to be able to convert output files to PMML?



Edited 1 time(s). Last edit at 09/13/2018 08:30AM by webmasterphilfv.

Options: ReplyQuote
Re: generate PMML output
Posted by: DIEP DAO
Date: September 14, 2018 11:44AM

Hi Philippe and Dang,

Thanks a lot for the reply. As a user of AR algorithms, I am very interested in visualizing the rules, particularly when there are so many of them. There exist some visualization packages for this. One is the arulesViz in R that I am using - accepting PMML file if visualizing rules output by external programs which are different from the arules package. From my experience with rule visualization, XML format seems widely accepted by many software.

Dang: I appreciate if you could share your code.

Many thanks to both of you again!

Diep

Options: ReplyQuote
Re: generate PMML output
Posted by: Dang Nguyen
Date: September 17, 2018 02:39PM

Hi Diep,

You can download the code from https://1drv.ms/t/s!Aoe-n4O4gxSgwSFcTh2895BNYIdZ

To visualize the rules, one way is to use the arules package in R as your said or another way is to use the Gephi software to create interactive graphs which are much more flexible than those generated by arules. An example of rule visualization can be found in my paper: https://www.sciencedirect.com/science/article/pii/S0950705118303824

If you need more help, please let me know.

Best,
Dang Nguyen

Options: ReplyQuote
Re: generate PMML output
Posted by: Diep Dao
Date: September 18, 2018 05:28AM

Hi Dang,

Thank you so much for the code and the suggestion on using Gephi. I had a great time reading your paper - what an interesting work!

Happy mining!

Diep

Options: ReplyQuote


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