Machine Learning-Based Assessment and Optimization of Electrode Materials for Supercapacitors Srikanta Moharana, Bibhuti B. Sahu, Jayakishan Meher, Rozalin Nayak, Ram Naresh Mahaling, Karthik Dhandapani, Kalim Deshmukh Nanostructured Materials for Energy Storage I IV, 2024 In the past few years, supercapacitors (SCs) have received remarkable attention in the field of energy conversion and storage systems in sustainable nanotechnologies because of their unique properties such as better cycling life, low processing cost, and excellent specific power than that of batteries. The carbon material, metal oxide, and conductive polymers are three electrode materials that have been utilized for processing the SCs owing to their superior electrochemical performances. However, machine learning (ML) techniques have placed great significance on quickly developing applications in a number of domains, including physics, chemistry, and material science, where vast amounts of data are available, to estimate the capacitance of electrode materials for SCs. In order to discover and tailor the energy storage material for electrochemical performance, ML plays an imperative role and also optimizes the properties of electrode materials in terms of high specific capacitance for SC applications. This chapter is intended to portray the contemporary technological applications, fundamental key concepts, and prospective domain of ML, in order to predict the electrode materials for design, development, and applications of SCs. Also, it will provide a platform for in-depth insight into ML to study the relationship between various parameters of electrode materials linked to SC performance, such as charge–discharge cycle, power, energy density, and specific surface area.
Disease detection in infected plant leaf by computational method A. K. Rath, J. K. Meher Archives of Phytopathology and Plant Protection, 2019 Plant diseases can directly affect the production hampering the economy significantly. Thus, early and correct detection of the disease is always a priority for an agriculture-dependent state. Of the many modern techniques of early detection of plant diseases, image processing has become a potential tool through which not only the disease can be detected early and correctly, but also it can be quantified successfully. The detection of two most important diseases of rice i.e., brown spot and rice blast was done through efficient computational method using a low complexity radial basis function neural network (RBFNN) classifier. The performance was analyzed using quality measures viz. accuracy, precision and recall and found to be 95%, 97% and 95%, respectively.
New encoded single-indicator sequences based on physico-chemical parameters for efficient exon identification J.K. Meher, P.K. Meher, G.N. Dash, M.K. Raval International Journal of Bioinformatics Research and Applications, 2012 The first step in gene identification problem based on genomic signal processing is to convert character strings into numerical sequences. These numerical sequences are then analysed spectrally or using digital filtering techniques for the period-3 peaks, which are present in exons (coding areas) and absent in introns (non-coding areas). In this paper, we have shown that single-indicator sequences can be generated by encoding schemes based on physico-chemical properties. Two new methods are proposed for generating single-indicator sequences based on hydration energy and dipole moments. The proposed methods produce high peak at exon locations and effectively suppress false exons (intron regions having greater peak than exon regions) resulting in high discriminating factor, sensitivity and specificity.
