Automatic classification of chemical behaviour by sequential hypothesization, multiparametric curve-fitting and deviation-pattern recognition-I Fully computerized evaluation of the functionality of a base from titration-curve data.

A completely automatic procedure is described for evaluating the functionality j of a base from the potentiometric titration curve obtained on titration with standard acid. It involves successive hypotheses that j = 1, 2,..., fits to the corresponding titration-curve equations, and scrutiny of the resulting deviation patterns. Logical decisions are based on both the existence and position of a region in which the deviations are much larger than the estimated standard error of measurement.