How to prune useless rules?

[bpdot]

antecedents	consequents	antecedent support	consequent support	support	confidence	lift	leverage	conviction	antecedent_len
(banana)	(eggs)	0.054217	0.427711	0.054217	1.0	2.338028	0.031028	inf	1
(banana, apple)	(eggs)	0.054217	0.427711	0.054217	1.0	2.338028	0.031028	inf	2

The "(banana, apple) -> (eggs)" rule doesn't improve confidence compare to "(banana) -> (eggs)" rule.
So I want to prune "(banana, apple) -> (eggs)" rule.
How can I do that?

[rasbt]

There is no specific API for pruning. You can use the pandas API on the resulting data frame to remove individual rows. E.g.,

selection = (df['antecedents'] == frozenset({'banana', 'apple'})) & (df['consequents'] == frozenset({'eggs'}))
new_df = df.loc[ ~selection ]

[bpdot]

Thanks for your reply.
There are many useless rules to be pruned. So filtering them one by one may not be acceptable.

[rasbt]

Yeah, having some automated pruning might be a good enhancement

[rasbt]

Added it to the issue tracker at #795