Combining Heuristics and Learning for Entity Linking

Entity linking refers to the task of mapping name strings in a text to their corresponding entities in a given knowledge base. It is an essential component in natural language understanding applications, but a challenging task. This paper proposes a method that combines heuristics and learning for entity linking by (i) learning coherence among co-occurrence entities within the text based on Wikipedia’s link structure and (ii) exploiting some heuristics based on the contexts and coreference relations among name strings. The experiment results on TAC-KBP2011 dataset show that our method achieves performance comparable to the state-of-the-art methods. The results also show that the proposed model is simple because of using a classifier trained on just two popular features in combination with some heuristics, but effective.