Natural Language Processing (NLP) has long been used to extract information from large bodies of text. NLP is often used to intelligently parse large volumes of data where the manual alternative may be infeasible. Named Entity Recognition (NER) is used to extract named entities such as people, places or organizations from text written in natural language. Using NER, NLP algorithms can be created to extract the mentions of geographic locations of different types from current and archived news articles. This information can be used to add a spatial window into previously flat datasets, allowing users to access information by filtering location information. Information that is derived can be used to support intelligent decision making and influence expert systems. This paper describes the development of an algorithm that uses the principles of both NLP and NER to extract references to geographic locations within news articles. The algorithm has been developed using the NLTK and Pattern Web Toolkit for Python and performs with a precision and accuracy above eighty (80) percent.
Wright, Jean-Mark and Mansingh, Gunjan, "An Algorithm to Extract Jamaican Geographic Locations from News Articles – Using NLP Techniques" (2014). CONF-IRM 2014 Proceedings. 24.