Semantic Technology by Xin Wang & Francesca A. Lisi & Guohui Xiao & Elena Botoeva

Author:Xin Wang & Francesca A. Lisi & Guohui Xiao & Elena Botoeva
Language: eng
Format: epub
ISBN: 9789811534126
Publisher: Springer Singapore

2 Related Works

There have been several works to address the problem of hypernymy detection, which can be categorized into path-based approaches and distributional approaches. A path-based approach identifies the hypernymy relationship through a lexical-syntactic path that joins the occurrence of pairs of entities in a large corpus. Snow’s work [8] learned to detect hypernymy relationship by a neural network model. The limitation of this approach is caused by the data sparsity. It usually requires co-occurrence of term pairs in a sentence, which can not be fulfilled in many scenarios. In recent years, the distributional approach focus on supervised methods, in which entity pairs are represented by a feature vector, i.e. the contexts with which each entity occurs separately in the corpus. In these methods, a classifier is trained on these vectors to predict hypernymy. Different from previous approaches, we construct a Sentence Graph from a Chinese traffic legal corpus, which is used to extract the local and global context to solve the problem of data sparsity.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.