A multidisciplinary approach that involves language universals, linguistic discourse analysis and computer information technology are combined to support the descriptive nature of this research dissertation. Utilising comparative methods to determine rudimentary language structures which reflect both the scientific and historic parameters that are embedded in all languages. From a hypothesis to the proof of concept, a multitude of computer applications have been used to test these language models, templates and frameworks. To encapsulate this entire approach, it is best described as "designing then building the theoretical, experimental, and practical projects that form the structural network of the Maori language system". The focus on methods for integrating the language is to investigate shared characteristics between Maori and New Zealand English. This has provided a complete methodology for a bilingual based system. A system with text and speech for language generation and classification. This approach has looked at existing computational linguistic and information processing techniques for the analysis of each language�s phenomena; where data from basic units to higher-order linguistic knowledge has been analysed in terms of their characteristics for similar and/or dissimilar features. The notion that some language units can have similar acoustic sounds, structures or even meanings in other languages is plausible. How these are identified was the key concept to building an integrated language system. This research has permitted further examination into developing a new series of phonological and lexical self organising maps of Maori. Using phoneme and word maps spatially organised around lower to higher order concepts such as �sounds like�. To facilitate the high demands placed on very large data stores, the further development of the speech database management system containing phonological, phonetic, lexical, semantic, and other language frameworks was also developed. This database has helped to examine how effectively Maori has been fully integrated into an existing English framework. The bilingual system will allow full interaction with a computer-based speech architecture. This will contribute to the existing knowledge being constructed by the many different disciplines associated with languages; naturally or artificially derived. Evolving connectionist systems are new tools that are trained in an unsupervised manner to be both adaptable and flexible. This hybrid approach is an improvement on past methods in the development of more effective and efficient ways for solving applied problems for speech data analysis, classification, rule extraction, information retrieval and knowledge acquisition. A preliminary study will apply bilingual data to an �evolving clustering method� algorithm that returns a structure containing acoustic clusters plotted using visualisation techniques. In the true practical sense, the complete bilingual system has had a bi-directional approach. Both languages have undergone similar data analysis, language modelling, data access, text and speech processing, and human-computer network interface interaction.
| Identifer | oai:union.ndltd.org:ADTP/217597 |
| Date | January 2001 |
| Creators | Laws, Mark R., n/a |
| Publisher | University of Otago. Department of Information Science |
| Source Sets | Australiasian Digital Theses Program |
| Language | English |
| Detected Language | English |
| Rights | http://policy01.otago.ac.nz/policies/FMPro?-db=policies.fm&-format=viewpolicy.html&-lay=viewpolicy&-sortfield=Title&Type=Academic&-recid=33025&-find), Copyright Mark R. Laws |
Page generated in 0.0016 seconds