Fostering Students’ Idea Generation through Corpus-based Recommendation in Online Writing Environment

Abstract

Idea generation plays an important role in the writing process. The use of natural language techniques to mine semantic relationships between vocabularies or sentences in order to provide students with more suitable contents for writing ideas. Therefore, this study developed a corpus-based Chinese writing recommendation system. This study collects, analyzes, processes, and constructs a primary school corpus with 580,000 written texts, and develops a vocabulary and sentence recommendation mechanism for writing. 37 fourth-year students are invited to participate in the evaluation. The study found that students provided 60% of the candidate vocabularies or sentences they provided, but 40% of the candidate vocabularies or sentences were not applicable. Furthermore, students had more than 60% positive attitude towards the system.