t-SNE
In this experiment you are presented with a graphical representation of thousands of bird songs. By touching a point on the map you can hear the song of the corresponding bird. The position and proximity of the points to each other tell something about the similarity of the bird songs. To achieve such a representation, a statistical method called t-SNE was used.
How does t-SNE work?
t-SNE stands for "t-distributed Stochastic Neighbor Embedding". steht für "t-distributed Stochastic Neighbor Embedding". It is a mathematical dimensionality reduction technique that transforms large amounts of data so that it can be represented on a flat plane while keeping similar data points close to each other
What is special about t-SNE is that it is largely possible to retain the relationships between points in the original data, even when they are shown in this flat representation. Bird songs can have thousands of frequency dimensions, but t-SNE is able to represent them in only two dimensions.
Why is this useful?
By simplifying high-dimensional data into a two-dimensional view, the data can be visually represented and patterns, groups, or anomalies can be more easily identified.
Where does t-SNE find application?
t-SNE is used, among other things, in the genome analysis of individual cells to understand how different cell types differ and how they interact with each other.
Despite careful control of the content, we assume no liability for the content of external links. The operators of the linked pages are solely responsible for their content.