This paper describes the application of several data mining approaches to solve a calibration problem in a quantitative chemistry environment. Experimental data obtained from reactions which involve known concentrations of two or more components are used to calibrate a model that, later, will be used to predict the (unknown) concentrations of those components in a new reaction. This problem can be seen as a selection + prediction one, where the goal is to obtain good values for the variables to predict while minimizing the number of the input variables needed, taking a small subset of really significant ones. Initial approaches to the problem were principal components analysis and filtering. Then we used methods to make smarter reductions of variables, by means of parallel estimation of distribution algorithms (EDAs) to choose collections of variables that yield models with less average prediction errors. A final step was to use multiobjective parallel EDAs, in order to present a set of optimal solutions instead of a single solution.