Background
System Overview and Goals
Morphological AnalysisSeeks to select a segmentation metric for isiZulu morphological analysis algorithms through statistical analysis, that scores analyzers based on a performance metric.
Natural Language GenerationImplement a library that models the grammatical rules of isiZulu and generates natural language which can develop language learning exercises.
Difficulty Evaluation and CompositionCompose balanced isiZulu language learning exercises, in accordance with Bloom's Taxonomy, which can have its responses automatically graded and provide feedback.
Conclusions and Results
Morphological AnalysisProduced a statistical analysis of the performance of isiZulu morphological analyzers on separate analysis tasks. Analysed corpus data for use in future Natural Language Processing tasks.
Natural Language GenerationSuccessful development of the isiZulu grammar and natural language generation library as well as successful use of the library to generate language learning exercises.
Difficulty Evaluation and CompositionAverage score obtained for a question decreases and average question complexity increases as one progresses through the levels of Bloom’s taxonomy.
Exercise Configuration Tool
Exercise Presentation
Exercise Feedback
The Team
Kgotso NkosiMorphological AnalysisNKSKGO001@myuct.ac.za
Nikhil GilbertNatural Language GenerationGLBNIK001@myuct.ac.za
Soham SinghDifficulty Evaluation and CompositionSNGSOH004@myuct.ac.za
Dr. Maria KeetSupervisorDepartment of Computer Science, University of Cape Townmaria.keet@uct.ac.za
