Term of Award

Fall 2022

Degree Name

Master of Science in Computer Science (M.S.)

Document Type and Release Option

Thesis (restricted to Georgia Southern)

Copyright Statement / License for Reuse

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Department

Department of Computer Science

Committee Chair

Andrew Allen

Committee Member 1

Ryan Florin

Committee Member 2

Lixin Li

Abstract

Mushrooms are a day-to-day ingredient that we cook with, use for medicine, and even sometimes use in tea. However, what happens when a mushroom turns out to be harmful? Poisonous mushrooms have caused 52 deaths and 704 exposures to harm throughout history and an average of 7 death in the U.S per year. It is hard to differentiate if a mushroom is poisonous or edible, and many people get easily confused. This in the end leads to these deaths and harmful exposures. Mushrooms have many attributes about them, 23 within the dataset used. Without studying them for years, it is very difficult to identify them at first glance. This paper aims to research and differentiate between poisonous and edible mushrooms and then find the best classification algorithm to classify them. Three of the most popular algorithms in machine learning are used to classify the dataset of mushrooms: decision trees, random forests, and neural network algorithms. These supervised and deep learning algorithms are going to be used to analyze a dataset of over 8000 various mushrooms, which are then going to be compared through accuracy scores. All three of the machine learning algorithms have undergone data preprocessing through label encoding to prevent bias. Then all three algorithms were trained on 80 percent of the data and then tested on the other 20 percent to get raw accuracy scores to see which is the most efficient. The algorithms will compute and analyze all 23 attributes with variations of those unique attributes to classify the mushrooms as edible or poisonous to evaluate which algorithm is the most efficient and accurate.

OCLC Number

1361718962

Research Data and Supplementary Material

No

Share

COinS