IN SILICO MODEL QSPR FOR PREDICTION OF STABILITY CONSTANTS OF METAL-THIOSEMICARBAZONE COMPLEXES
In the present work, the stability constants logb<sub>11</sub> and the concentration of metal ion and thiosemicarbazone in complex solutions were determined by using <em>in silico</em> models. The 2D, 3D, physicochemical and quantum descriptors of complexes were generated from the molecular geometric structure and semi-empirical quantum calculation PM7 and PM7/sparkle. The quantitative structure and property relationships (QSPRs) were constructed by using the ordinary linear regression (OLR) and artificial neural network (ANN). The best linear model QSPR<sub>OLR</sub> (with <em>k</em> of 6) involved descriptors k0, core-core repulsion, xp5, xch5, valence, and SHHBd. The quality of model QSPR<sub>OLR</sub> had the statistical values: <em>R</em><sup>2</sup><sub>train</sub> = 0.898, <em>R</em><sup>2</sup><sub>adj</sub> = 0.889, <em>Q</em><sup>2</sup><em><sub>LOO</sub></em> = 0.846, MSE = 1.136, and <em>F<sub>stat</sub></em> = 91.348. The neural network model QSPR<sub>ANN</sub> with architecture I(6)-HL(6)-O(1) had the statistical values: <em>R</em><sup>2</sup><em><sub>train</sub></em> = 0.9768, and <em>Q</em><sup>2</sup><em><sub>LOO</sub></em> = 0.8687. The predictability of QSPR models for complexes of the test group turned out to be in good agreement with those from the experimental data in the literature.