Type

Conference Proceedings

Authors

Catherine Mooney
Brett A. Becker

Subjects

Computer Science

Topics
compiler error messages error message categorization novice programmers compiler errors principal component analysis cs1 error analysis java programming

Categorizing Compiler Error Messages with Principal Component Analysis (2016)

Abstract Being a competent programmer is critical for students in all computing disciplines and software engineering in particular. Novice programming students face a number of challenges and these have been shown to contribute to worrying dropout rates for students majoring in computing, and the growing number of non-majors who are learning to program. Methods of identifying and helping at-risk programming students have been researched for decades. Much of this research focuses on categorizing the errors that novice programmers make, in order to help understand why these errors are made, with the goal of helping them overcome these errors quickly, or avoid them altogether. This paper presents the first known work on categorizing compiler errors using Principal Component Analysis. In this, we find a new way of discovering categories of related errors from data produced by the students in the course of their programming activity. This method may be used to identify where these students are struggling and provide direction in efforts to help them.
Collections Ireland -> University College Dublin -> School of Computer Science
Ireland -> University College Dublin -> Computer Science Research Collection

Full list of authors on original publication

Catherine Mooney, Brett A. Becker

Experts in our system

1
Catherine Mooney
University College Dublin
Total Publications: 63
 
2
Brett A. Becker
University College Dublin
Total Publications: 20