Contents
Do you know how to use t-SNE effectively?
We’ll walk through a series of simple examples to illustrate what t-SNE diagrams can and cannot show. The t-SNE technique really is useful—but only if you know how to interpret it. Before diving in: if you haven’t encountered t-SNE before, here’s what you need to know about the math behind it.
What can you do with the t-SNE algorithm?
The principal components (pca1, pca2, etc.) can be used as features in classification or clustering algorithms. Now you will do the same exercise using the t-SNE algorithm. Scikit-learn has an implementation of t-SNE available, and you can check its documentation here.
How is t-SNE used to analyze flow cytometry?
Conversely, t-SNE assembles closely-related data points to neighbor each other in space, increasing the ability to resolve differences between data points that are quite similar. An important effect of this, however, is that physical distances between clusters on a t-SNE map do not indicate how closely or distantly related that data is.
Why are the points spread randomly in t-SNE?
That helps distinguish neighbor’s probabilities and because you’ve already understood the whole process you should be able to adjust it to new values. The next part of t-SNE is to create low-dimensional space with the same number of points as in the original space. Points should be spread randomly on a new space.
What should the learning rate be for t-SNE?
The learning rate for t-SNE is usually in the range [10.0, 1000.0]. If the learning rate is too high, the data may look like a ‘ball’ with any point approximately equidistant from its nearest neighbours. If the learning rate is too low, most points may look compressed in a dense cloud with few outliers.
Which is a tuneable parameter in the t-SNE?
Those differences can be a major source of confusion. A second feature of t-SNE is a tuneable parameter, “perplexity,” which says (loosely) how to balance attention between local and global aspects of your data. The parameter is, in a sense, a guess about the number of close neighbors each point has.