Polynomial Graph Degree Classification

Table of Contents

• Introduction
• Technologies
• Installation
• Usage
• Results

Introduction

This project aims to classify polynomial graphs based on their degrees using Convolutional Neural Networks (CNNs). It contains code for generating graphs, as well as for training and evaluating the machine learning models.

Technologies

Graph Generation (graph_generation.py)

• Python Libraries: NumPy, Matplotlib, Pandas, OpenCV
• Methods: Graphs are generated using NumPy for mathematical calculations and Matplotlib for plotting.
• Data Storage: Graph images and metadata are saved to disk.

Model Training (model_training.py)

• Python Libraries: NumPy, Pandas, Matplotlib, TensorFlow, scikit-learn
• Methods: The machine learning model uses a Convolutional Neural Network (CNN) implemented in TensorFlow. Data is split into training and testing sets using scikit-learn.
• Data Loading: Data is loaded from a pickled dataset.

Installation

Clone the repository and navigate to the project directory. Install the required packages using pip:

git clone https://github.com/WilliamHackspeare/Graph-Degree-Classification.git
cd Graph-Degree-Classification
pip install -r requirements.txt

Usage

Graph Generation

Run graph_generation.py to generate polynomial graphs.

python graph_generation.py

Model Training

Run model_training.py to train the machine learning model.

python model_training.py

Results

Model Performance

Training & Validation Loss

Training & Validation Accuracy

Test Loss & Accuracy

• Test Loss: 0.20
• Test Accuracy: 0.94

Test Predictions