DEMAx Tool Based on an Improved Model for Semiautomatic C/C++ Source Code Assessment
Date Issued
2021-04-16
Author(s)
Angelovski, Damjan
DOI
10.1145/3470716.3470728
Abstract
As the demand for software engineers rises, so does the demand for
their education. With the increasing number of students, educators
struggle to keep up.We aim to ease their burden by providing a new
tool for semiautomatic source code assessment, named DEMAx. It
analyzes C/C++ source codes and their test case results and with the
help of machine learning, provides information on the likelihood
that a submission should be manually assessed.
In this paper we present a tool with the focus on the new improvements
of our previous work that include direct static analysis
of non-compiling code and ranking metrics of the source codes.
At the end, we present the results of the improved model on the
testing data, which are solid ground for the use of our tool.
their education. With the increasing number of students, educators
struggle to keep up.We aim to ease their burden by providing a new
tool for semiautomatic source code assessment, named DEMAx. It
analyzes C/C++ source codes and their test case results and with the
help of machine learning, provides information on the likelihood
that a submission should be manually assessed.
In this paper we present a tool with the focus on the new improvements
of our previous work that include direct static analysis
of non-compiling code and ranking metrics of the source codes.
At the end, we present the results of the improved model on the
testing data, which are solid ground for the use of our tool.