4_1_Stock_et_al.pdf (235.72 kB)
He Tatai Whenua: Automated Extraction of Landscape Terms and their Meanings in New Zealand Maori. GeoComputation 2019
Version 2 2019-12-01, 23:23
Version 1 2019-09-18, 00:50
conference contributionposted on 2019-12-01, 23:23 authored by Kristin Stock, Hone Morris, Margaret Forster, Rex Paraku, Ekaterina Egorova
Current geographic information systems largely represent Western conceptualisations of landscape, and indigenous worldviews are rarely incorporated, resulting in decision making that does not consider all perspectives. The incorporation of New Zealand Maori worldviews in such systems requires a better understanding of the ways that Maori think about land. We use corpus linguistics to address this problem, demonstrating the automated extraction and analysis of content from early newspaper editions in the Maori language using word co-occurrences, which we then explore using an analytical frame developed for New Zealand Maori. We also demonstrate a method for automatically classifying geographic senses of landscape terms in the Maori using machine learning with a bag of words approach with SVM, a particular challenge due to the high degree of polysemy in the language.