Extracting Spatio-Temporal Information from Chinese Archaeological Site Text

Archaeological site text is the main carrier of archaeological data at present, which contains rich information. How to efficiently extract useful knowledge from massive unstructured texts great significance for mining and reuse According information (such as name, location, cultural type, dynasty, etc.) recorded in Chinese text, this paper combines deep learning natural language processing techniques study extraction method automatically obtaining spatio-temporal sites. The initial construction corpus completed first time, input into Bidirectional Long Short-Term Memory with Conditional Random Fields (BiLSTM-CRF) entity recognition model Gated Recurrent Units Dual Attention (BiGRU-Dual Attention) relationship training. F1 values BiLSTM-CRF BiGRU-Dual on test set reach 87.87% 88.05%, respectively. demonstrates that proposed feasible texts, promotes establishment graphs archaeology provides new methods ideas development technology archaeology.