The results obtained by minimizing the sum of squares of different functions of measured variables x, t, y, P; with and without weighting are compared. The unweighted regression leads to different values of estimated parameters according to the type of minimization criterion. It is shown that the weighted regression of x, t, y data gives rise to the same values of estimated parameters not depending on the type of minimization criterion. This is also true for x, t, P data. It is shown that the simultaneous correlation of x, t, y, P data is erroneous but this statement is valid only for "classical" treatment. A very simple method is proposed based on the concept of "multiresponse model" for treatment of all (x, t, y, P) data together.
ASJC Scopus subject areas
- Chemical Engineering(all)
- Industrial and Manufacturing Engineering