NER-based military simulation scenario development process

Abstract

For combat simulation, the simulation scenario serves as the foundation and data source. It is not, however, easy to develop military simulation scenario because the developing process of these texts was time-consuming. To solve this problem, in this paper, we propose a distant supervised method for developing military simulation scenarios based on named entity recognition (NER) method. This method consists of three phases: extracting the key elements of simulation scenario, recognizing named entities of the text, and generating an executable simulation scenario. First, we analyze the two types of scenarios involved in the development process of military simulation scenarios: operational scenario and executable scenario. Second, we train a NER model on operational scenario corpus. Then, we compare our distant supervised-based NER method with the other NER methods, and we achieve an overall improvement of F1 score of 9.01%. Finally, to demonstrate the feasibility of our approach, we use a case study to implement a combat simulation scenario development progress.

Chinese Named Entity Recognition in Football Based on ALBERT-BiLSTM Model

“…Named entity recognition in the Chinese language is used in agriculture [39], natural hazards [40], the military [41], engineering [42], chemicals [43], and mainly in medicine, covering electronic health records [43,44] and clinical texts [45,46]. Although named entity recognition has many applications in Chinese, it still has a variety of challenges.…”