Crosstabs, Support, Confidence and Lift

"People who buy X also buy Y" analysis starts with a crosstab. For two boolean (or categorical) columns, the association-rule measures are:

• support of A ∧ B: the fraction of rows where both hold
• confidence A → B: P(B | A)
• lift: P(A ∧ B) / (P(A) · P(B)). Above 1 means they occur together more often than chance.

Implement:

• crosstab(df, row, col, normalize=None): pd.crosstab with normalize as None, "index", "columns" or "all"
• rule_metrics(df, a_col, a_val, b_col, b_val): a dict with support, confidence and lift for the rule "a_col == a_val → b_col == b_val"
• top_rules(baskets, min_support): baskets is a list of item sets. For every ordered pair of distinct items (A, B) with support ≥ min_support, compute the metrics, and return a list of (A, B, lift) sorted by lift descending, then A, then B.