A Software System for Automatic Construction of a Semantic Word Network
Дмитрий Алексеевич Усталов, Андрей Владимирович Созыкин · Bulletin of the South Ural State University Series Computational Mathematics and Software Engineering · 2017
A semantic word network is a network that represents the semantic relations between individual words or their lexical senses. In this paper, we present a software system for automatic construction of a semantic word network. The system, called SWN, is designed for the construction for such a semantic word network and includes the implementation of unsupervised concept discovery and semantic relation establishing methods as well as the implementation of a supervised relation expansion method. The methods use widely available language resources, such as semantic relation dictionaries and background text corpora. The domain model has been presented using the VOWL notation. The system architecture is represented using the UML notation and is composed of the concept discovery module, semantic relation construction module, the Semantic Web export module, and the evaluation dataset construction module based on microtask-based crowdsourcing. The present software system is open source and is available for integration into third-party data mining systems.