Get Inspire insights from former attendees in our AMA discussion thread on Inspire Buzz. ACEs and other community members are on call all week to answer!

Alteryx Designer Desktop Discussions

Find answers, ask questions, and share expertise about Alteryx Designer Desktop and Intelligence Suite.

MB Analysis - Lift

Daniel_He
5 - Atom

Hello together,

 

I am trying to do a market basket analysis.

Therefore, I used the MB Rule- & MB Inspect tool. The workflows works just fine.

However, the "Lift" KPI is calculated incorrectly.

 

From testing with a small sample size I think I figured out that Alteryx gives out the "Support" of the LHS item/item bundle as the "Lift" of the whole LHS & RHS-rule (see red marker).

If I calculate "Lift" manually as [Support productbundle AB] / ([support product A] * [support product B]), I do get different results than Alteryx. E.g. for Egg -> Salt, Lift should be 2. However, Alteryx gives out 0,5 as the corresponding Lift (just the same as the support value..., see yellow marker).

Daniel_He_1-1631783243382.png

 

 

I am desperatly trying to find my/the error. Can someone help me? Find my simple testing workflow and the input-Excel attached.

 

Thanks in advance!

 

Best regards,

Daniel

4 REPLIES 4
robbinrv
8 - Asteroid

Hi Daniel, bit late to the party here but just noticed the same issue on our workflows here.
On version 2020.4 I get the correct expected lifts but when opened in 2021.2 I get the wrong Lift (just the LHS support).
Dummy workflow attached as well.
Feels like something is wrong in the way the report/data is produced/output.

Created a support ticket, let's see if they can help

shiqi
5 - Atom

Hi all, I am encountering the same issue on v2021.3. Has a fix been found?

vishnuakshay
5 - Atom

Hi All, I am having same issue on 2021.3 version.can any one help me with this.?

jzamora
7 - Meteor

In 2021.3 Ateryx is returning the Coverage metric instead of the Lift metric. This is a very disappointing error.

This function in R (using arules package) 

DATAFRAME(apriori(yourdata, parameter = list(minlen=2, supp=0.002, conf=0.05)))

returns a column with support, confidence, coverage, lift and count

 

jzamora_0-1662437579714.png

 

Labels