|
Title:
|
AUTOMATIC EVALUATION OF PROGRAMMING USING STATISTICS TECHNIQUE: AN APPROACH TO COMPLEMENT TEST-BASED EVALUATION |
|
Author(s):
|
Mireille Pinheiro Moreira , Eloi Luiz Favero , Thaís Lira Tavares Dos Santos |
|
ISBN:
|
978-972-8924-83-6 |
|
Editors:
|
Miguel Baptista Nunes and Maggie McPherson (series editors: Piet Kommers, Pedro Isaías and Nian-Shing Chen) |
|
Year:
|
2009 |
|
Edition:
|
V I, 2 |
|
Keywords:
|
Automatic evaluation, multiple linear regression, teaching of programming |
|
Type:
|
Full Paper |
|
First Page:
|
296 |
|
Last Page:
|
303 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
This paper presents a new approach for automatic evaluation of programming. This approach proposes the usage of the
statistical technique called multiple linear regression as a complement to test-based evaluation of algorithms and
programming. The usage of this technique allows the students to verify their answers - even though they are correct -
looking for a better new solution. The multiple linear regression uses a set of indicators, the software engineering metrics,
that allow us to know how complex is the student's program in terms of: number of declarations, number of functions,
and so on. This approach has been developed and applied in an e-learning environment for teaching programming, and
the final accuracy for this technique was more than 90%. |
|
|
|
|
|
|
