Voltammetric Electrode Array Optimization for Black Tea Discrimination Using Computational Intelligence Approach

Abstract

In this paper, a computational intelligence approach has been adopted to optimize the voltammetric electrode array for discrimination of the overall quality of black tea. Electrochemical measurements have been performed using the three- electrode system that comprises of an Ag/AgCl reference electrode, a platinum counter electrode, and synthesized working electrodes (WE). Nine working electrodes (WE1 to WE9) have been synthesized by varying the compositions of polymer and graphite. The electrodes have been imbibed in black tea samples and subjected to triangular voltage waveform ranging from -0.04 V to 0.8 V. From the data set so obtained, the number of electrodes have been optimized using computational approaches followed by a polling technique. Features were extracted by four feature transformation techniques and samples were classified with five different classification methods. Polling score has been assigned to each WE based on the decisions obtained by different feature transformation and classification techniques. The average classification accuracy rate of two working electrodes (WE1 and WE8) optimized from the algorithms were 91.27% and 91.34%, respectively. Thus, the optimization technique implemented in the present work yielded acceptable results and this technique could be suitable for other real time applications as well.