Straightforward method to predict the impact sensitivity of nitro aromatic compounds using molecular images

In the present work, multivariate image analysis combined with quantitative structure property relationships was considered to predict impact sensitivity of nitro aromatic energetic compounds. The analysis was performed between 2D chemical structures and impact sensitivity for nitro aromatic energetic compounds for the first time. Principal component analysis ranking-adaptive neuro-fuzzy inference systems (PCA Ranking-ANFIS) was employed to investigate relationship existed between descriptors and impact sensitivity. The analysis also statistically validated for its predictive power using external validation set, cross-validation, applicability domain and Y-randomization evaluation procedures. The satisfactory results (R2p=0.907, Q2LOO=0.895, R2L25%O=0.850, RMSELOO=0.152, RMSEL25%O=0.185 and r2m=0.711) make clear that the proposed model can be used to predict the impact sensitivity of new nitro aromatic compounds for engineering. The proposed analysis can be used to predict the impact sensitivity of new nitro aromatic energetic compounds.

Keywords:

Impact sensitivity , Nitro aromatic compounds , MIA-QSPR , Adaptive Neuro-Fuzzy Inference Systems