Neuro-Fuzzification Architecture for Modeling of Electrochemical Ion-Sensing Data of Imidazole-Dicarboxylate-Based Ru(II)–Bipyridine Complex

Abstract

Anion-and pH-sensing behaviors of an imidazole-dicarboxylatebased Ru(II)−bipyridine complex possessing a number of dissociable protons in its secondary coordination sphere are employed here for the creation of multiple Boolean and fuzzy logic systems. The absorption, emission, and electrochemical behaviors of the metalloreceptor were significantly modulated upon the influence of basic anions (such as F − , AcO − , and H 2 PO 4 − ) as well as by altering the pH of the solution. Interestingly, the deprotonation of the metalloreceptor by selected anions or by alkaline pH, followed by its restoration to its original form by acid or acidic pH is reversible and could be repeated many times. The metalloreceptor is capable to demonstrate several advanced Boolean functions, namely, three-input OR gate, set−reset flip-flop logic, and traffic signal, by employing its electrochemical responses through proper use of different inputs. Administering exhaustive sensing experiments by changing the analyte concentration within a wide range is usually tedious as well as exorbitantly costly. To get rid of these difficulties, we employed here several soft computing approaches such as artificial neural networks (ANN), fuzzy logic systems (FLS), or adaptive neuro−fuzzy inference system (ANFIS) to foresee the experimental sensing data and to appropriately model the protonation−deprotonation behaviors of the metalloreceptor. Reasonably good correlation between the experimental and model output data is also reflected in their tested root-mean-square error values (0.115961 and 0.118894 for the ANFIS model).

Harnessing Deep Neural Networks to Analyze Multi‐Channel Anion Sensing Characteristics of a Ru(II)‐Pyrazolyl‐Bis(Benzimidazole) Complex

“…We previously reported multi‐channel anion‐ and pH sensing behaviours of an imidazole‐dicarboxylate based Ru(II)‐bipyridine complex possessing a dissociable imidazole NH and a carboxylic acid proton in its secondary coordination sphere [76] . In a separate work, anion‐ and temperature responsive behaviours of three luminescent Ru(II)‐terpyridine complexes possessing varying number of imidazole NH protons (1 to 3) in their outer coordination sphere were also investigated [77] .…”

Neural network and decision tree-based machine learning tools to analyse the anion-responsive behaviours of emissive Ru(ii)–terpyridine complexes

“…Recurrent (RNN) and function fitting (ANN-FF) networks are frequently used. [31][32][33][34][35][36][37][38][39] In this present work, ANN-FF procedure has been employed due to the static nature of our system and because of its ability to understand and forecast complex systems. Although ANN is quite competent in grasping the data but not so competent to understand the meaning of every neuron and its load.…”

Analysis and prediction of anion- and temperature responsive behaviours of luminescent Ru(ii)-terpyridine complexes by using Boolean, fuzzy logic, artificial neural network and adapted neuro fuzzy inference models

“…Comparatively little progress has been accomplished in other supplementary areas of AI, such as FL, ANNs, ANFIS and evolutionary computation. [10][11][12][13][14][15][16][17][18][19][20] Creation of a dependable and exhaustive catalogue might expand ML to a wide range of applications. Hence, tremendous efforts have now been paid to expand the productive domain of AI having vague and imprecise inputs.…”

“…Among the two types of neural networks {recurrent (RNN) and feedforward (FFN)}, we applied here the ANN-function fitting (ANN-FF) network due to the static nature of our systems and because of its efficiency in understanding and forecasting a complicated system. [17][18][19][20] It is evident that neural networks (NNs) possess an efficient grasping ability of the data but not so efficient in terms of understanding the meaning of each neuron and its weight. In contrast to NNs, the FL-based model could be understood better as it operates via linguistic languages and the IF-THEN law.…”

“…Among the two types of neural networks {recurrent (RNN) and feed-forward (FFN)}, we applied here the ANN-function fitting (ANN-FF) network due to the static nature of our systems and because of its efficiency in understanding and forecasting a complicated system. 17–20…”