FEATURE-SELECTION AND VALIDATION OF SIMCA MODELS - A CASE-STUDY WITH A TYPICAL ITALIAN CHEESE

A strategy for building class models by means of SIMCA (soft independe nt modelling of class analogy) is suggested. to be applied in the case of a small number of objects and a large number of variables. The str ategy uses both the customary procedure, based on the selection of var iables with both high modelling and discrimination powers, and a novel procedure. Here Monte Carlo simulations are used to obtain the signif icance level of the experimental Fisher weights, so that only the rele vant variables are selected, avoiding the use of noisy information. Th e validation of SIMCA models is performed by means of a leave-one-out procedure: many validation parameters are suggested to evaluate the ac curacy of the models obtained. Data on a typical Italian cheese have b een used to show the feature selection and the validation procedures. The significance of the validation parameters has been tested by compa ring the results of the 'cheese categories' with those obtained from a rtificial and real data sets (the variety Versicolor of the iris flowe r, and categories of typical wines and olive oils). The models compute d for the typical cheese are shown to be reliable.