A comparison of two approaches for hint generation in programming tutors (abstract only)

  • Authors:
  • Andrew Hicks;Barry Peddycord II;Irena Rindos;Christopher Simmons

  • Affiliations:
  • North Carolina State University, Raleigh, NC, USA;North Carolina State University, Raleigh, NC, USA;North Carolina State University, Raleigh, NC, USA;North Carolina State University, Raleigh, NC, USA

  • Venue:
  • Proceedings of the 45th ACM technical symposium on Computer science education
  • Year:
  • 2014

Quantified Score

Hi-index 0.00

Visualization

Abstract

Intelligent tutoring systems have been shown to significantly aid in student learning without requiring extensive instructor intervention. One feature of intelligent tutors is their ability to provide hints to students who get stuck, but the production of hints by hand is prohibitively expensive. Hint Factory (Stamper et al., 2008) is a domain-independent technique for generating hints from interaction graphs. In this work, we explore approaches for creating graphs that represent interactions in an educational programming game called BOTS. We use data from seven programming puzzles that contain a total of 1100 interactions with the game. We believe that we can begin by seeing what works in the small language of a programming game to inform how we can scale the techniques for "real" languages in "real" tutors.