Hunger Hacker: An Exploration of Neural Networks for Food Preference Prediction

Name: Joseph Henry
Major: Computer Science
Minor: Communication Studies
Advisor: Drew Guarnera; Thomas Montelione (second reader)

This study explores what makes up neural networks and the subsequent creation and design of the application “Hunger Hacker”. This application utilizes a neural network to predict what type of food a user wants to eat at that moment. The goal of this application is to solve the age old question many are presented with everyday, “What do I want to eat?” This question is explored by diving into the basics of neural networks, problems neural networks face, activation functions, feature selection, feature extraction, and neural network design theory. A proposed neural network is then presented that predicts what type of food someone wants. This neural network is then showcased and built into an application so that it can easily receive user input. This work is important because it reveals how accurate neural networks can be at solving humanity’s everyday questions, and if the process can be carried out for applicable issues. The findings of the study showcase the issues that arise when trying to create a neural network for such a complex problem, and that with a little more work and adjustments, this process can be carried out on a large scale for these types of problems.