Towards an Automatic Scoring for Code Assessment in the Universities: A Case Study the Algerian University

The assessment of learners, whether conducted in traditional classrooms or in the Computer Environment for Human Learning (CEHL), has consistently been a subject of ambiguity and lack of clarity for both the evaluated and the evaluator on the one hand, and between the evaluators themselves on the other hand. The approach taken during the validation of the results of the automated assessment tools was to make an analytical comparison with the scores given by the human experts. The goal of this research paper is to compare and contrast two versions of the S_Onto_ALPPWA automated assessment scoring tool with the expert’s scoring. To achieve the underlined goal, we used the correlational research methodology to demonstrate the extent of the correlation between S_Onto_ALPPWA and the manual scores provided by the experts. Our obtained results indicate that, in the majority of instances, the S_Onto_ALPPWA scores are closely aligned with the experts’ scores. In our research, we used Ward’s hierarchical clustering techniques to categorize the experts into three distinct groups: lenient, severe, and moderate experts. This approach was employed to gather relevant information.