Affiliation:
1. Amity Institute of Pharmacy, Amity University, Noida, India
2. B.S. Anangpuria Institute of Pharmacy, Haryana, India
Abstract
Background:
In this study, computational Artificial Neural Network (ANN) model is
applied for optimisation and evaluation of silver nanoparticles (AgNPs) size in the bionanocomposite
matrix. The primary purpose of this study is used a feed-forward ANN model to
create a connection between the output as the size of Ag–NPs, with four inputs variables, including
AgNO3 concentration, the weight percentage of starch, Bentonite amount and Gallic acid concentration.
Method:
Silver nanoparticles were synthesised via biogenic green reduction method. The fast Levenberg–
Marquardt (LM) backpropagation algorithm applied for the training of ANN model in this
research. The optimised ANN is a multilayer perceptron (MLP) which is a kind of feed forward (4-
10-1) network has an input layer with 4 nodes, hidden layers with 10 neurones, and an output layer
with 1 node found a fitness function.
Results:
The output results of developed computational ANN model were compared to its predictive
values of the size of silver nanoparticles regarding two statistical parameters, the coefficient of determination
(R2) and mean square error (MSE) of data set. It observed that ANN predicted values are
close to the actual values and well fitted to the data. The mean square error(MSE) is 0.03, and a regression
is about 1.
Conclusion:
AgNO3 concentration has the most likely factor affecting the size of silver nanoparticles
(Ag–NPs) and this makes possible to develop a green reduction method for the preparation of silver
nanoparticles. This study confirms that employing ANN method with LM feed forward (4-10-1)
network is a useful tool with cost-effective for predicting the results of analysis and modelling of the
chemical reactions.
Publisher
Bentham Science Publishers Ltd.
Subject
Pharmaceutical Science,Biomedical Engineering,Medicine (miscellaneous),Bioengineering,Biotechnology
Cited by
16 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献