COMPARISON OF 3 DIFFERENT APPROACHES TO THE PROPERTY PREDICTION-PROBLEM

Neighborhood subspace approximation method has been developed for solv ing the property prediction problem. In this paper the performance of this method is evaluated on three groups of compounds. The molecular s tructure was encoded as sequences of well-known structural (topologica l) indices. Euclidean distance has been used for determining the simil arities between compounds. Property prediction results of the new meth od are compared with the results of the neighborhood based approximati on (clustering) method and with the results obtained by an artificial intelligence program using machine learning tools.