learning rate, perplexity) really matterĬluster sizes in a t-SNE plot mean nothingĭistances between clusters might not mean anythingįor topology, you may need more than one plot This Distill paper How to Use t-SNE Effectively gives a great summary of the common pitfalls of t-SNE analysis. In other words, tuning learning rate and perplexity, it is possible to obtain very different looking 2-d plots for the same number of training steps and using the same data. The original paper states: “The performance of t-SNE is fairly robust to changes in the perplexity, and typical values are between 5 and 50.” However, my experience is that getting the most from t-SNE may mean analyzing multiple plots with different perplexities. In this context, perplexity itself is a stab in the dark on how many close neighbours each observation may have and is user-provided. However, you can adjust it to do so by tuning the perplexity parameter, which regulates (loosely) how to balance attention between local and global aspects of your data. t-SNE does not necessarily focus on the local structure. Having said that, I would be doubly careful about making inferences on a dataset just looking at t-SNE plots. There are also radial basis function networks, which I am not an expert on. You have listed one of them: SVM's with RBF, and has listed kNN. In terms of classification, this immediately brings to mind instance-based learning methods.
Specifically, the probabilities are generally given by a normalized Gaussian kernel computed from the input data or from the embedding. SNE techniques compute an N ×N similarity matrix in both the original data space and in the low-dimensional embedding space in such a way that the similaritiesįorm a probability distribution over pairs of objects.
First a brief answer, and then a longer comment:
0 kommentar(er)
