Biometric Data Profiling With Machine Learning

Closed Human-Computer Interaction Computer Science Master Project Thesis No Affiliation

Goals

• Creating data sets with movement data of different persons in various scenarios
• Application and analysis of different machine learning methods on such data sets

Tasks

We will work out the exact tasks together, and your own ideas are welcome. In general, however, the project will most likely include one or more of the following tasks:

• Conducting studies to collect movement data
• Analysis and visualisation of movement data
• Researching the effects of data selection and pre-processing on machine learning models
• Researching the effectiveness of different machine learning methods

Prerequisites

• Experience with Machine Learning (e.g. HCI Machine Learning course): you should have already done a project with Machine Learning, terms like “Overfitting”, “Gradient Descent” or “Confusion Matrix” shouldn’t be news to you
• Programming experience: you will most certainly have to do a lot of programming, so you should be comfortable with coding, StackOverflow and git.
• Basic knowledge of Python, Numpy and Pandas: you’ll do data analysis and machine learning, for that you’ll use Python and it will help you if you already know your way around Numpy and Pandas

Contact Persons at the University Würzburg