Simpler Machine Learning Using Spreadsheets: Neural Network Predict

  • Leong Thin-Yin Singapore University of Social Sciences (SUSS)
  • Leong Yonghui Jonathan

Abstract

Machine Learning as a phenomenon has gone viral, with many technologists and software vendors promoting it. However, offered tools remain highly technical and not accessible to those without rigorous training in Computer Science or Business Analytics. It would be more useful if end-users can understand it beyond the sales pitch or blind application, and perhaps, even work from scratch to build simple models without much additional training. With better assimilation and acceptance of this AI methodology as an acquired skill and not just head knowledge, many more may want to invest the intensive effort to learn the required tough mathematics and cryptic programming. Or, after simple trial explorations, be willing to put aside substantial budgets to employ skilled professionals for full-scale business application. With simplicity and accessibility in mind, this paper renders Neural Network, a key machine learning methodology, on the ubiquitous and easily comprehensible spreadsheet without macros or add-ins, employing only elementary operations and if so desired, optionally leveraging on its built-in Solver. We will show that backpropagation can be achieved using the elegant though obscure recursive computation feature, with no need for Solver. We will demonstrate the application of neural network on a familiar problem: early and prior prediction of students’ graduation GPA. The paper can be used to form the core content for introducing machine learning to non-technical audiences, particularly those majoring in Business and the Social Sciences.
Published
Feb 21, 2020
How to Cite
THIN-YIN, Leong; JONATHAN, Leong Yonghui. Simpler Machine Learning Using Spreadsheets: Neural Network Predict. European Journal of Engineering and Formal Sciences, [S.l.], v. 4, n. 1, p. 124-138, feb. 2020. ISSN 2601-6311. Available at: <http://journals.euser.org/index.php/ejef/article/view/4638>. Date accessed: 03 aug. 2020. doi: http://dx.doi.org/10.26417/ejef.v4i1.p124-138.