attributecommonly

Attributecommonly is a term used in information science to describe attributes that are widely present across a dataset or population. It denotes properties that occur with high frequency among entities and are often used to guide data modeling, schema design, and inference. While not a formal standard, attributecommonly serves as a practical concept for identifying core attributes that characterize a class of objects.

Formally, consider a set of entities E and for each entity e a set of attributes A(e). Given a threshold τ in [0,1], an attribute a is attributecommonly if the proportion of entities in E that possess a is at least τ. In more complex cases, attributecommonly can be defined at class or subclass levels, or vary over time or context, yielding dynamic or hierarchical variants.

Applications include prioritizing schema design around core attributes, guiding indexing and storage decisions, improving data quality through consistent attribute definitions, and informing feature selection in machine learning. For example, in a customer information system, attributes such as customer_id, email, and geographic region often qualify as attributecommonly across most records, while niche fields may not.

Variants of the concept address context dependence, such as time-bound attributecommonly (attributes frequently present within a time window) or hierarchical attributecommonly (core attributes at different levels of a taxonomy). The idea relates to, but is distinct from, frequent itemsets and core attributes used in data profiling and knowledge representation.

See also: frequent itemset, core attribute, data profiling, feature selection.