```xml <result> <column_name>precision_at_k</column_name> <family>additive</family> <subtype>intensive_additive</subtype> <confidence>0.85</confidence> <requires_user_confirmation>false</requires_user_confirmation> <sum_meaningful>false</sum_meaningful> <signed>false</signed> <reducers> <primary_reducer>mean</primary_reducer> <valid_reducers>mean, min, max, median, distribution</valid_reducers> <excluded_reducers>sum (summing precision scores is meaningless), product (not a ratio of sequential changes)</excluded_reducers> <average_semantics>arithmetic_mean</average_semantics> </reducers> <inner_outer_measure> <inner_measure_reducer>min</inner_measure_reducer> <outer_measure_reducer>max</outer_measure_reducer> </inner_outer_measure> <multiplicative_partner_derivable>false</multiplicative_partner_derivable> <reasoning> Gate prior: DIMENSION table with no sequence axis makes multiplicative highly implausible. The column name "precision_at_k" is a machine learning evaluation metric representing the fraction of top-k predictions that are correct - a bounded ratio in [0,1] but NOT a sequential growth factor. Values centered near 1.0 (mean=0.92, median=1.0, range [0.75,1.0]) initially suggest multiplicative, but this is a LEVEL metric per entity, not a compounding ratio. On a DIMENSION table, these are snapshot quality scores where averaging across entities is meaningful but summing is not. Classified as intensive_additive (a level/intensity) rather than ratio_multiplicative because there is no sequence to compound along and the metric represents absolute performance at a point, not relative change. </reasoning> <table_type_conflict>false</table_type_conflict> </result> ```