TSIR: A Chinese Temporal semantics Information Retrieval system based on MapReduce

The significance of time in information production and consumption has been recognised in information retrieval research. Temporal information plays an important role in the webpage retrieval. The webpage has both the temporal metadata and temporal semantics in the content. However, the existing search engines conduct the information retrieval based on text keywords rather than temporal semantics. To address this issue, a Temporal semantics Information Retrieval (TSIR) System is proposed to deal with the Chinese temporal information retrieval. The TSIR system is deployed on Hadoop and implemented by the means of MapReduce. Firstly, the Chinese temporal regular expression rule is introduced to extract the explicit and implicit temporal phrases in the query keywords and webpages. Secondly, the scores of webpages are re-evaluated by taking text relevance and temporal semantics relevance into account and the returned results are ranked according to re-evaluation. Experiment shows that TSIR system could precisely and effectively match the keywords queries related to temporal expression.