Prediction and optimizing of methylene blue sequestration to activated charcoal/magnetic nanocomposites using artificial neutral network and response surface methodology

Performance Evaluation of Benzyl Alcohol Oxidation with tert-Butyl Hydroperoxide to Benzaldehyde Using the Response Surface Methodology, Artificial Neural Network, and Adaptive Neuro-Fuzzy Inference System Model

“…Considering the statistical results obtained in this work, the ANN model was the least efficient in terms of precision and accuracy in estimating the benzaldehyde yield, while the ANFIS model was better than the ANN model. The structure where the error value approaches zero (0) and the superlative R 2 -value is near 1 is the acceptable exceptional model performance . Therefore, the error analysis revealed the superiority of ANFIS at simulating the oxidation process of benzyl alcohol to benzaldehyde over RSM and ANN techniques.…”

“…However, the observed high values of R , R 2 , and adjusted R 2 are indicative of the good fit of the models . For a model to be adjudged reliable and acceptable, its R 2 value must be greater than 0.8. , Consequently, the adjusted R 2 was used to check the overestimation of the R 2 . The obtained adjusted R 2 values were sufficiently very close to R 2 , which further validated the correctness and accuracy of the predicted models .…”

“…However, the design consists of 2 k factorial factors, which define the number of factorial points, enhanced by 2 k axial factors, which were chosen at a distance of α from the design center, and center factors, which help reduce the experimental error, as some experimental repetitions are necessary. , In addition, the factorial points showed equal variations between the low and high values (+1 and −1), while the axial points were chosen to be (α = ±1.68) to ensure that the orthogonality and rotatability of each independent variable are measured at five levels of (−α, −1, 0, 1, +α). ,− Therefore, based on what we want to achieve, the range of values, actual and coded, were inputted into the statistical tool in such a way that their minima, central, and maxima values take the values of (−α, 0, and +α) (Table ). The differences in axial points for each level divided by four give the incrementing value for each independent variable.…”

“…The structure where the error value approaches zero (0) and the superlative R 2 -value is near 1 is the acceptable exceptional model performance. 53 Therefore, the error analysis revealed the superiority of ANFIS at simulating the process of benzyl alcohol to benzaldehyde over RSM and ANN techniques. The actual experimental values as well as the predicted values of benzaldehyde yield for the three models were plotted against the run numbers (Figure 9b).…”