This study examines the mutual relatedness of 318 major topological indices (TIs) for three sets of molecules: (i) a set of 139 hydrocarbons, (ii) a diverse set of 1029 compounds and (iii) a diverse set of 2887 compounds. The TIs included in this study are those that have been frequently used in the characterization of structure and QSAR/QSPR studies. After variable reduction based on the elimination of TIs for which all values were zero and those that were completely correlated with another TI, a variable clustering technique was used to cluster the TIs which resulted in 16, 37 and 56 clusters, respectively, for the three data sets mentioned above. Analysis of the correspondence among the clusters derived from the three groups of chemicals has been carried out in an effort to understand the dimensionality of the structure spaces derived for the three different sets of chemicals and the structural aspects characterized by the various TIs.
|Original language||English (US)|
|Number of pages||14|
|Journal||Croatica Chemica Acta|
|State||Published - May 1 2004|
- Cluster analysis
- Diverse compounds databases
- Topological indices