Artificial neural networks applied for studying metallic complexes

Abstract

Metallic complexes of multimetal and multiligand systems are complicated for calculating equilibrium concentrations in solutions. An artificial neural network has been developed for studying Al 3+ and EDTA complexes in solution with an initial concentration of 0.01 mol L −1 for these species. In this system there are 20 compounds and may exist 18 simultaneous reactions. The neural network has been trained and the simulated data of different concentrations as a function of pH are predicted with an accuracy of about 1% for all species simultaneously. A general analytical formula is presented, which directly relates all the concentrations as a function of pH. The analysis showed that predictions closer to the boundary of the input and output data are quantitative while out of these limits these are not even qualitative.

Analysis of the applicability of artificial neural networks for studying blood plasma: determination of magnesium ion concentration as a case study

“…The bias index number refers to the neurons of an artificial neuron of a certain layer, for example b 1, 1 to 5 indicates the 5 bias (from 1 to 5) of each artificial neuron of the first hidden layer. metal ions in other chemical systems has been reported and satisfactory results have been found (21)(22)(23).…”

An Alternative Approach Based on Artificial Neural Networks to Study Controlled Drug Release

“…These weights can be determined in a very efficient way, which will be discussed later. In general, the weights are corrected as5,24 where $\Delta W_{ji}^l$ represents the correction to the weight between the j th element in the l th layer and i th element in the previous layer. The quantity ${\rm out}_i^{l - 1}$ contains the output result of the l −1 layer.…”

“…These quantities determine the rate of convergence during the training procedure in which they are dynamically adjusted to obtain the best rate of convergence during the training procedure in which they are dynamically adjusted to obtain the best rate of convergence. The errors determined during the training stage are given by5,24 and where y j is the output target, which is compared with the output ANN results $({\rm out}_j^l )$ of the l th layer. In this work, the neuron behavior was calculated with the sigmoid function for the intermediary layer and a linear function for the output layer 23,24…”

“…Accordingly, alternative approaches should be studied that combine the advantages of conventional methods already established with those of other techniques, such as artificial neural networks (ANNs). The latter methodology has been applied, for example, to study metallic complexation5 and to simulate the drug release profile in drug delivery systems 6–15…”

A search map for organic additives and solvents applicable in high-voltage rechargeable batteries

“…Moreover, if large data sets are available, the ANN can predict the physical and chemical properties of organic compounds quantitatively, which has been a challenge for chemists. 20,21 Indeed, the ANN and machine learning methods have been applied to the research areas of drug design, 22,23 metallic complexes, 24 sunset yellow dye adsorption, 25 ionic liquids, 26,27 and inorganic materials, [28][29][30] etc.…”