Date of Award

Spring 4-26-2019

Document Type

Honors Project

Degree Name

Bachelor of Science

Department

Mathematics

Department Chair or Program Director

Randall Helmstutler, Ph.D.

First Advisor

James Collins, Ph.D.

Second Advisor

Julius Esunge, Ph.D.

Third Advisor

Jangwoon Lee, Ph.D.

Major or Concentration

Mathematics

Abstract

The purpose of this research is to decrease the run time of Bertini, a program that approximates solutions of polynomial systems. Bertini can be run more efficiently if it is known whether a polynomial is singular or non-singular. In this research, we focus on polynomials in one variable. We use a machine learning algorithm to classify polynomials into these two categories. To do so, we create and use a set of polynomials to train a neural network and create a model. Then, we create and use a test set to assess the accuracy of the model. By changing the hyper-parameters of the system and by changing the functions used in the system, the accuracy of the model is able to be increased.

Included in

Mathematics Commons

Share

COinS