alsoconnections

Alsoconnections is a term used in network science and data modeling to denote a secondary form of connection between entities, added in addition to primary relation types. They capture alternative or supplementary associations that are not captured by direct links, such as shared attributes, co-participation, or functional similarity.

In a knowledge graph, alsoconnections may be represented as edges labeled also or as a separate relation category with distinct semantics. They can be undirected or directed, weighted to indicate strength, and are often derived from data mining of co-occurrence patterns, attribute similarities, or transitive closure of related items. They serve to enrich connectivity without redefining core relationships.

Uses of alsoconnections include enriching recommender systems, where these links indicate alternatives to a user’s preferred items; social networks to suggest users with parallel interests; and bibliographic databases to connect works with similar topics beyond direct citations. They can also support query expansion and semantic search by offering related entities that extend primary results.

Limitations and considerations involve maintaining interpretability and avoiding graph “noise,” since excessive alsoconnections can obscure primary relations. Best practices include thresholding by strength, temporal consideration to reflect changing similarities, and allowing explicit user control over the weighting of alsoconnections. They are related to, but distinct from, synonyms, co-occurrence edges, and inference rules in multi-relational graphs, and are typically evaluated in terms of added predictive value and preserved explainability.