Interest Measure Package for Association Rules and Class Association Rules

This package provides a set of interest measures for association rules and class association

Installation

pip install interest-measures

Usage

from interest_measures import InterestMeasures

im = InterestMeasures(A=[0.1, 0.2, 0.3, 0.4],
                      B=[0.5, 0.6, 0.7, 0.8],
                      AB=[0.1, 0.2, 0.3, 0.2], N=100)

print(im.confidence)  # [1. 1. 1. 0.5]
print(im.lift)  # [2. 1.66666667 1.42857143 0.625 ]
print(im.conviction)  # [inf inf inf 0.4]

Available Interest Measures

• Accuracy
• Added value
• Chi square
• Collective strength
• Complement class support
• Conditional entropy
• Confidence
• Confidence causal
• Confirm causal
• Confirm descriptive
• Confirmed confidence causal
• Correlation coefficient
• Cosine
• Coverage
• Dir
• F measure
• Gini index
• Goodman kruskal
• Implication index
• J measure
• Kappa
• Klosgen
• K-measure
• Kulczynski 2
• Least contradiction
• Leverage
• Lift
• Loevinger
• Logical necessity
• Mutual information
• Normalized mutual information
• Odd multiplier
• Odds ratio
• One way support
• Piatetsky Shapiro
• Prevalence
• Putative causal dependency
• Recall
• Relative risk
• Specificity
• Support
• Theil Uncertainty Coefficiente
• Tic
• Two way support