FITTING MODELS TO DATA: THE MATHEMATISING STEP IN THE MODELLING PROCESS

Abstract

This paper presents a mathematical modelling activity experienced with students of first year university level centred on a problem of forecasting sales using onevariable functions. It then focuses on the back and forth movements between the initial system – a time-series of the term sales of a firm – and the different models proposed to make the forecasting. The analysis of these movements, that are at the core of the ‘mathematising step’ of the modelling cycle, shows how the initial empirical system is being enlarged and progressively enriched with new variables and mathematical objects. Thus the development of a modelling activity initiated with a real-situation may soon lead to a process where the mathematising affects both the system and the model. 1. THE MATHEMATISING STEP IN THE MODELLING PROCESS