What is the impact of specifying a positive value in FP-Growth Parameter settings

batstache611batstache611 Member Posts: 45 Guru
edited December 2018 in Help



As per operator documentations, the positive value parameter of the FP-Growth operator is defined as follows -


"positive value

Description: This parameter determines which value of the binominal attributes should be treated as positive. The attributes with that value are considered as part of a transaction. If left blank, the ExampleSet determines which value is used."

I wish to know what the second part means - "If left blank,....". How does the exampleset determine which values are used? My dataset is comprised of semi-structured text-mined data and I'm trying to look for word associations. The output of Binominal to Polynominal operator gives me a word vector with true/false values. Should I still specify the positive value at FP growth? Thank you.



Best Answer

  • Options
    IngoRMIngoRM Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, Community Manager, RMResearcher, Member, University Professor Posts: 1,751 RM Founder
    Solution Accepted



    Yes, I would do it.  The reason is that in specific cases you might actually want to look for associations with "false", i.e. with things which are NOT happening.  So it is always better to specify it.  If you do not specify it, the first value in the internal value mapping of the example set it considered to be the positive (if I remember correctly).  But this is something you typically do not want to depend on since this order might change with different data loading or preparations.


    Hope this helps,


Sign In or Register to comment.