جستجوی مقالات مرتبط با کلیدواژه « Multilinear Regression » در نشریات گروه « شیمی »

There is a need for knowledge, experience, laboratory, materials, and time to conduct chemical experiments. The results depend on the process and are also quite costly. For economic and rapid results, chemical processes can be modeled by utilizing data obtained in the past. In this paper, an artificial neural network model is proposed for predicting the removal efficiency of Cr (VI) from aqueous solutions with Amberlite IRA-96 resin, as being independent of chemical processes. Multiple linear regression, linear and quadratic particle swarm optimization are also used to compare prediction success. A total of 34 experimental data were used for training and validation of the model. pH, amount of resin, contact time, and concentration were used as input data. The removal efficiency is considered as output data for each model. The statistical methods of root-mean-square error, mean absolute percentage error, variance absolute relative error and the coefficient of determination were used to evaluate the performance of the developed models. The system has been analyzed using a feature selection method to assess the influence of input parameters on the sorption efficiency. The most significant factor was found in pH. The obtained results show that the proposed ANN model is more reliable than the other models for estimating removal efficiency.

The Quantitative Structure–Activity Relationship of a series of novel Thiazoline derivatives with anticancer activity has been studied by using the density functional theory by B3LYP/ 6-31G. Descriptors of quantum mechanics of 21 thiazoline derivatives with known activity were obtained. Multiple linear regressions were employed to model the relationships between molecular descriptors and biological activity of molecules using stepwise method. The most model shows not only significant statistical quality, but also predictive ability, with the square of adjusted correlation coefficient (R2=0.945) and standard error (SE=0.586). We find that the anticancer activity expressed that as half maximal inhibitory concentration (IC50), closely relates to the highest occupied molecular orbital, dipole moment, softness, hardness, ionization energy, electron affinity. Accordingly can be offered a quantitative model, and interpret the activity of the compounds relying on the multivariate statistical analysis. This study shows that the prediction results were in excellent agreement with the experimental value.  The results can offer some useful references for understanding the action mechanism and designing new compounds with anticancer activity.