Vector-based Ranking Techniques for Identifying the Topical Anchors of a Context

Terms in textual documents tend to occur more in contexts to which they are related. We can exploit this bias by modelling terms and their cooccurrences s, resembling the abstract semantic memory structures in human beings. Given a context in the cooccurrence graph, we identify the topical anchors of the context. A context is defined as a collection of terms and topical anchors are those terms whose semantics represent the topic of the whole context. In graph-theoretic terms, a topical anchor is the node which is central to the context subgraph. In this work we propose different vector-based ranking techniques and rank the nodes in a subgraph to identify topical anchors. Experiments were conducted on 100 distinct human defined contexts. The topical anchors generated were evaluated against those chosen manually by 86 volunteers and the results show that the algorithm correctly identifies the topical anchors 70% of the time.

• Type: Conference
• Year: 2009
• Where: COMAD
• Authors: Aditya Ramana Rachakonda, Srinath Srinivasa
• Keywords: Abstract Semantic Memory, COMAD 2008, COMAD 2009, Context Subgraph, Topical Anchors