New Algorithms for Smart Assessment of Math Exercises

This paper deals with the field of mathematics education where the aim is to generate exercises with randomized data. The process of exercise generation involves, first, identification of common errors that may be performed when solving the exercise, then, modeling of these errors by appropriate functions and recognition and distinction of errors through algorithms. Thus, although randomized, the exercise parameters must be chosen in such a way that it would be possible to understand the error the student committed and with it to guarantee a proper feedback on his answer. We call this process smart assessment. In this paper we present two algorithms for exercise generation admitting smart assessment.