To Extend or Not to Extend? Context-Specific Corpus Enrichment

摘要

An agent in pursuit of a task may work with a corpus of documents with linked subjective content descriptions. Faced with a new document, an agent has to decide whether to include that document in its corpus or not. Basing the decision on only words, topics, or entities, has shown to not lead to a balanced performance for varying documents. Therefore, this paper presents an approach for an agent to decide if a new document adds value to its existing corpus by combining texts and content descriptions. Furthermore, an agent can use the approach as a starting point for high quality content descriptions for new documents. A case study shows the effectiveness of our approach given varying types of new documents.