Fig. 2: The framework of the proposed position-sensitive word vector training model. | Humanities and Social Sciences Communications