search

rasbt / mlxtend

A library of extension and helper modules for Python's data analysis and machine learning libraries.
https://rasbt.github.io/mlxtend/
Other
4.85k stars 857 forks source link

Incorrect Documentation about Support for Association Rules #849

Closed BrunoBarre closed 1 year ago

BrunoBarre commented 2 years ago

Hello,

There seems to be an incoherence in the support metric description for association rules in this page http://rasbt.github.io/mlxtend/user_guide/frequent_patterns/association_rules/#example-3-frequent-itemsets-with-incomplete-antecedent-and-consequent-information.

The formula is correct : support(A→C)=support(A∪C),range: [0,1]. But in the following paragraph, it is written The 'support' metric then computes the support of the combined itemset A ∪ C -- note that 'support' depends on 'antecedent support' and 'consequent support' via min('antecedent support', 'consequent support'). Which is clearly different from the formula.

In mlxtend.frequent_patterns.association_rules support is computed using the formula, and not the minimum between the antecedent support and consequent support.

zuari1993 commented 1 year ago

Hi, Thanks for pointing this out. You are correct on the inconsistency. I will work on fixing this.