Розроблення просторово-зваженого критерію геоприв’язки природномовних текстів до масивів вод

Abstract

У роботі розглядається проблема автоматичної геоприв’язки природомовних текстів до масивів вод.

The paper considers the problem of automatic georeferencing of natural language texts to water bodies. Previous studies have shown the need to take into account the spatial aspect of geographical names to improve the accuracy of the work. For this purpose, a new spatially weighted criterion has been developed based on the adaptation of the TF-IDF method to spatial data referencing (Multiplicity-Length Term with Inverse Spatial Object Frequency). Thus, we combine the assessment of the uniqueness of spatial geonames and their spatial extent with specific water bodies. Due to the complex analysis, the accuracy of the automatic referencing process is significantly increased.