Moreover, it is found that ANN can model the process accurately, and is able to predict the model outputs very close to those predicted by ASPEN HYSYS simulator for refined palm oil process.
The model output corresponded very well with the measured E. coli and enterococci concentrations in the river; the Root Mean Square Error and the Nash-Sutcliffe efficiency for Log10-transformed concentrations were found to be 0.23 (Log10 CFU/100 ml) and 0.84 for E. coli, and 0.19 (Log10 CFU/100 ml) and 0.86 for enterococci, respectively.
Numerical results demonstrate that the ETA can effectively solve the (ell_{1}) regularized model and output very sparse solutions of co-coercive NCPs with high quality.
The HI-2 is dB for the real-world output and dB for the model output, that illustrates a very good accuracy of the identification method.
The model output for indoor measurements is very similar to personal measurement model, shown in SI Table S3.
There is also the issue that a wide range of model parameters could give a very similar model output and these methods would have difficulty coping with this.
We acknowledge that our model outputs could be very different given another eruption location, sequence, duration (e.g., multiple sporadic phases), and eruptive volume.
For instance, the noisy interaction pairs from high throughput experiments in human could be used to build neighbor pairs for training SML model (e.g. embedding on output) very naturally and thus the method has significant potential for intra-species PPI predictions such as in human.
The GSA results of compactness and lacuna count both indicated that variation of the rigidity of the cells (λ A ) has very little impact on the model output.
Results showed very low discrepancies between the real experiment measurements and the model output which ensures the applicability of this method.
