PROJECT IDEA FOR FINAL YEAR STUDENTS

Error-Driven Foreign Language Learning

Learning a foreign language is painstaking. Foreign language learners with different background (different mother tongue and different level of proficiency, etc.) are prone to make different types of mistakes. In an error-driven foreign language learning framework, learner’s errors are identified and annotated from a large number of people into a database. This collection is known as learner corpus. Patterns of errors and association of errors with learners can be easily identified using the annotated corpus and data mining algorithms (as it is done with shopping basket analysis in e-commerce to predict who is likely to buy which products). It is possible to teach foreign language effectively by identifying error-patterns in a learner and presenting the most relevant learning materials based on the mistakes a learner makes and likely to make. In this project, students will be required to collect and annotate errors in Arabic Speaker’s English followed by subsequent error analysis using machine learning and data mining algorithms. The students will also develop a prototype to demonstrate the effectiveness of error driven learning. Strong background in AI, XML and programming is necessary.