Te Taupanga

Abstract

This paper outlines the enhancement of an existing Kapa Haka competition judging system, aimed at automating the moderation of judges' inputs. The existing judging system is vulnerable to unintentional human bias and misinterpretation of the competition rules. The primary objective is to ensure the fair and efficient assessment of competitions. Leveraging Artificial Intelligence (AI), a model was developed using historical competition data extracted from the Haka in the Park 2018 competition results. The model's development involved forming a dictionary of polarising phrases, in the Māori language, from the extracted data. The selection of phrases from the extracted data, involved data analysis and consultation with domain experts. The resulting model utilises the generated dictionary to conduct classification as a form of lexicon-based opinion extraction. The model is provided with natural language text written in Māori which critiques a particular performance. The classifications produced identify inputs requiring correction. The classifications have been evaluated with input provided by domain experts at the Wellington Māori Cultural Society. This evaluation ensures that the model can generalise beyond the Haka in the Park 2018 competition data. Kapa Haka holds significant cultural importance in Aotearoa New Zealand. This project indirectly supports the overall well-being of the Kapa Haka community, promoting its growth and vitality. Overall, this project aims to revolutionise the judging process for Kapa Haka competitions, enhancing fairness and efficiency while promoting sustainability and community well-being.