Search

Computational neural networks have been developed to classify and quantify nine organic vapors. The neural network analyses used data that consisted of the change in fluorescence from a sensor array that consisted of 19 fiber optics with immobilized dye in polymer matrices. Plots of change in fluorescence intensity versus time were measured as pulses of analyte were presented to the sensor array. Descriptors were calculated from the intensity vs time plots, and they were used to build neural network models that accurately classified and quantified each of the nine analytes. Most of the data were used to train the neural networks (training set members), some were used to assist termination of training (cross-validation set members), and some were used to validate the models (prediction set members). Classification rates approaching 100% were achieved for the training set data, and 90% of the members in the prediction set were correctly classified. In addition, 97% of the prediction set observations were assigned a correct relative concentration.

Numerical representations of structure-based features are used to estimate both the retention indexes and sweetnesses of a diverse set of industrially important fragrance compounds. Retention indexes measured on nonpolar as well as polar stationary phases are modeled with accuracies of 3.6% and 5.6% at the mean of the respective retention ranges. Similar success was achieved when the developed equations were applied to predict the retention indexes of external data set compounds. Finally, the implications of using strictly 2-D structural information versus incorporating geometrical information are explored and discussed. The intensity of sweetness attributed to each compound is quantitatively predicted using identical multiple linear regression techniques. Difficulties encountered in this portion of the study warranted a critique of the procedures used to gain access to the odor data. As a consequence, the limited control exerted over several experimental variables is questioned.

High-quality regression equations (R greater than 0.996) modeling the gas chromatographic retention indices of 115 odor-active compounds on stationary phases of different polarity are generated by using the ADAPT software system. Multiple linear regression techniques are used to describe the statistical relationship between the Kováts retention indices and structure-based molecular descriptors. These descriptors encode topological, geometrical, and electronic features of the molecules. The utility of several new descriptors encoding functions of partial atomic charge and solvent accessible surface area is demonstrated. Quantitative predictions of odor threshold values for a subset of these compounds containing the alcohol functional group are also calculated by using a similar methodology.