Beyond extraction accuracy: addressing the quality of geographical named entity through advanced recognition and correction models using a modified BERT framework
In the realm of geospatial services and applications, the accuracy of address information is of utmost importance.Traditional methods of data collection, being both labor-intensive and costly, have prompted researchers to turn to Volunteered Geographic Information (VGI) for the extraction of Geographical Named Entity (GNE).Notwithstanding, prior st