Identifying Weak Sentences in student drafts: A tutoring system

Abstract

The first draft of an undergraduate student thesis generally presents deficiencies, which must be polished with the help of the academic advisor to get an acceptable document. However, this task is repeated every time a student prepares his thesis, becoming extra time spent by the advisor. Our work seeks to help the student improve the writing, based on intelligent tutoring and natural language processing techniques. For the current study, we focus primarily on the conclusions section of a thesis. In this paper we present three tutoring system components: Identifying Weak Sentences, Classifying the Weak Sentences, Customizing Feedback to Students. Our system identifies weaknesses in sentences, such as the use of general instead of specific terms, or the absence of reflections and personal opinions. We provide initial models and their evaluations for each component. © Springer International Publishing Switzerland 2014.