الفهرس 
AbstractOnly 14 pages are availabe for public view
 |  | from | 118 |  |  |
| | | from | 118 | | |
Abstract
Our study aims to develop specialized machine learning (ML) models with simple, rapid, and accurate predictive power in the graphene-based composite sub- category of material-based supercapacitors (SCs) electrodes. Open-source ML package Scikit-learn was used to develop the ML models that can; give input values related to graphene-based electrodes; predict electrode performance with a simple process and high accuracy. Investigating large amounts of experimental data can develop ML models that predict specific capacitance performance, reducing the need for trial-and-error experiments. Appropriately, accurate predictions can lead researchers to where spending their resources would have the maximum probability of making advancements. It is of huge importance to capture all key input characteristics during the ML model development. Data from more than two hundred research papers was collected (See Appendix B) and analyzed to develop four different ML models. These ML models developed complex models that describe the relationship between the features chosen in order to predict specific capacitance. A large number of never- before-used input features was used in our attempt to fully capture all variables that may influence the specific capacitance of SCs.
The following chapters present our work:
Chapter 1:
Chapter one introduces the concepts of SCs as a promising type of energy storage and machine learning as a promising technique for gaining insights from data.
Chapter 2:
The second chapter lays down the groundwork of the methodology used in our study. The process used to develop the ML models is described.
Chapter 3:
In this chapter our published research paper is presented in full. The chapter contains an abstract for the research paper, introduction, methodology, results and discussion, and conclusion.