Fitting lines and curves to data, least squares method

Abstract

In scientific and technological work we often have to measure two quantities, one of which is subject to a certain amount of unpredictable variation, often called ‘scatter’ (e.g. the yield in some chemical reactions is frequently subject to scatter), whereas the other quantity can be determined beforehand exactly (e.g. the

temperature or the duration of a reaction can be determined exactly, but the yield often cannot). The problem then is to find the true relation between the two quantities (e.g. exactly how does the yield vary with temperature).