Visualizing Mountains as High- Dimensional Parameter Spaces

📄 Project Abstract

Visualizing high-dimensional data is an important field in Science and Computing. Highdimensional data can consist of a parameter space that can be thought of as an N-dimensional landscape, comprising of mountains and valleys in relation to some fitness of the data. Visualizing this landscape is not a trivial thing, and it is often necessary to identify points in the data where these 'mountains' are wide as well as high, representing optimal regions in the data. This project looks at why it can be a difficult task, and how it can be achieved by researching methods of high-dimensional visualization and identifying strengths and weaknesses in the current tools. By identifying and creating a suitable data set that can be visualized with such tools; an idea for a new tool was developed which uses some aspects of the current techniques. The new tool was assessed against a current tool in a user-study; feedback was mostly positive and showed that the new tool has good potential and in some c...