연관규칙(Association rules) 파이썬 구현하기 mlxtend.frequent_patterns.apriori

from mlxtend.frequent_patterns import apriori
from mlxtend.frequent_patterns import fpmax
from mlxtend.frequent_patterns import association_rules

itemsets = apriori(fp_df, min_support=0.2, use_colnames=True)
itemsets.sort_values("support", ascending=False)


#fpmax(fp_df, min_support=0.2, use_colnames=True)


rules = association_rules(itemsets, min_threshold=0.5)

antecedents : X   
consequent : Y   
antecedent support : 데이터에서 X의 출현빈도   
consequent support : 데이터에서 Y의 출현빈도   
support(지지도) : X+Y의 출현빈도   
confidence(신뢰도) : 조건부 확률 s(X+Y)/s(X)   
lift(향상도) : confidence / (antecedent support*consequent support) #약간의 정규화 의미   1미만>negatively, 1초과>positively correlation 1>independent
leverage : ??   
conviction : ?? lift랑 비례한다?