Learning Web Programming: Identifying Complex Topics from Student Discussion Forums and Lecture Slides

Designing and delivering web application courses for computing undergraduates is a challenging task. Lack of understanding of web concepts affects the students” interest in web application development. Therefore., faculty employ several traditional strategies including hands on exercises and labs as well as innovative strategies such as videos and discussion forums. However., due to the volume of the posts in the forums., the instructors find it challenging to attend to the students' challenges and focus on more challenging topics across the classroom. In this paper., we propose a text mining based solution to extract the questions and complex topics. We evaluated the solution on the year 2 Web Application course offered to the computing undergraduates. Our experiments show that logistic regression model performs better in classifying question posts and cosine similarity performs better in assigning the topic label to the posts. The findings are visually depicted and are useful to the faculty to identify the topics that requires more attention to improve students' learning.