A new model for semiautomatic student source code assessment

Abstract

Programming courses at university and high school level, and competitions in informatics (programming), often require fast assessment of the received programming tasks solutions. This problem is usually solved by the use of automated systems that check the produced output for some test cases for every solution. In this paper, we present a new model for semiautomatic student source codes assessment for a given programming task, based on our approach of representation of the program codes as vectors. It represents a human and computer collaborative effort. Our research on the use of these vectors in data mining analysis of the source codes, with the achieved excellent results on the number of correctly clustered items, is a solid foundation for the proposed model. At the end, we present the results of the preliminary testing of the model.