Altair RISE

A program to recognize and reward our most engaged community members

Nominate Yourself Now!

FP-growth on marketing data

Hello,

is it possible to apply FP-growth when the variables are polynomial? I am trying to apply the operator to this dataset, from the book Elements of Statistical Learning:

https://web.stanford.edu/~hastie/ElemStatLearn/

I am having issues because the variables are polynomial. Can FP-growth handle this data set?

Best,
Bernardo

Find more posts tagged with

AI Studio

FP-Growth

Accepted answers

YYH

You may not have to re-code the numbers. But string "3" in income will be equivalent to the string "3" in age...
Suppose in the beginning, you have the numerical attributes loaded for marketing data, you can convert the values from numerical to polynominal. I would prefer to replace the nominal values "1","2","3" with the string values defined by data dictionary with "Map" or "Replace" operators
e.g. in gender

 1-> Male
 2-> Female
https://web.stanford.edu/~hastie/ElemStatLearn/datasets/marketing.info.txt

All comments

YYH

Hi @bernardo_pagnon, which data you are using from ESL textbook? Are you able to share it here? Thanks!

bernardo_pagnon

I tried, but I guess it is not supported here.

You got to enter here

https://web.stanford.edu/~hastie/ElemStatLearn/

then click on "Data" (white button on the left), and the "Marketing", and then on the hyperlink "data".

Regards,

Bernardo

YYH

Thanks @bernardo_pagnon for pointing out the source data. Still I am not sure what you want to extract from the data. Did I get the wrong data (Link direct to the income survey data) with 14 demographic attributes?
https://web.stanford.edu/~hastie/ElemStatLearn/datasets/marketing.data
https://web.stanford.edu/~hastie/ElemStatLearn/datasets/marketing.info.txt

bernardo_pagnon

Yes, it is the correct one. My goal was to extract relationships between info available. I guess I identified the problem: different attributes use the number 1 through 5 for instance, to represent the different polynomial values at each attribute. But RM cannot differentiate two numbers "3" in two different attributes: they represent completely different things, but RM identify them as the same.

I guess a solution would be to work with the data and generate different numbers for each attribute:

Attribute 1: 1-9

Attribute 2: 10-11

and so on.

Best,

Bernardo

YYH

 1-> Male
 2-> Female
https://web.stanford.edu/~hastie/ElemStatLearn/datasets/marketing.info.txt

bernardo_pagnon

Interesting, will try that!

Thank you so much for your time and for your patience.

Regards,

Bernardo