Temporal Entity Tagging untuk Dokumen Bahasa Indonesia

Abstract

One of the processes in document pre-processing is providing identifier tags called Named Entity Tagging. This research aims to implement a temporal entity tagging in Indonesian language documents with rule-based method. Temporal expressions are marked with DATE tag and the temporal value of the expression. Temporal expression comprise explicit expression, implicit expression, and event expression. The rules were obtained by analyzing temporal expressions in 100 Indonesian documents. There are 40 derived rules implemented in this system over 341 temporal expressions in the document. The system could find all of the temporal expressions consisting of 150 explicit temporal expressions, 155 implicit temporal expressions, and 36 event expressions. Also, the evaluation was conducted on 34 other documents. The system can correctly identify as many as 122 of 129 temporal expression contained in the additional documents. The system has achieved an accuracy of 100% for the 100 initial documents and 94.57% for the 34 test documents. Thus, rule-based identification of temporal entities for Indonesian performs well.