Migraines are Debilitating

Anyone who knows someone or suffers from migraines themselves can attest to how debilitating this neurological disorder can be. The most common form of treatment for migraines are abortive medications that need to be taken at the very beginning of a migraine or else they will not work. Often, head pain can occur and you have no way of knowing if it's going to lead to a severe episode or if it will dissipate. Moreover, migraine medications are expensive and difficult on the body to take regularly so every migraineur weighs the pros and cons of whether to take medication right away. This project is my attempt at helping people understand their migraines and help them make an informed decision.

Android GUI

The first form this project will take is as an Android app. Android provides a huge amount of flexibility for developing and distributing apps that will allow me to share my project with whoever needs it. By having it available in a portable format with an easy to use interface, someone suffering with a migraine can easily access it anywhere and reduce the necessary screen time required as much as possible.

Each Migraineur is Unique

The core idea of this app is to use a decision tree algorithm to ensure only the triggers that matter to each user are used for prediction. Each user will have their own predictive algorithm based on their inputs. Another really great use for this algorithm is that it will help identify the most important triggers and filter out those that have less of an effect. It's very difficult to figure out which triggers are actually having an impact and which are simply a confirmation bias. This has fascinating implications for migraine research as well as the most important factors can be identified for the group as a whole.

Machine Learning

Initially, there will not be enough data to make a unique prediction for an individual user but as each new event is recorded, the algorithm will readjust itself to fit the new data. The longer you use the app, the more powerful and accurate it will become.

Contact

Sarah Connor is a MSc of Computer Science student at Birkbeck College, University of London. For more information, please contact me at sconno05@dcs.bbk.ac.uk