• Courses
  • Tutorials
  • Jobs
  • Practice
  • Contests
August 15, 2022 |940 Views
t-SNE in Machine Learning
  Share   Like
Description
Discussion

In this video, we will be discussing t-SNE from Machine Learning in detail

T-distributed Stochastic Neighbor Embedding (t-SNE)  is a nonlinear dimensionality reduction technique.
It is useful for embedding high-dimensional data for visualization in a low-dimensional environment of two or three dimensions. It converts multidimensional data into two or more observable dimensions for humans. 
The next time you work with high-dimensional data, you might need to draw fewer exploratory data analysis charts thanks to the t-SNE algorithms.

There are two primary steps in the t-SNE algorithm:
1) First, t-SNE builds a probability distribution across pairs of high-dimensional objects, assigned a higher probability while dissimilar points are assigned a lower probability.  

2) Second, t-SNE minimises the Kullback-Leibler divergence (KL divergence) between the two distributions with regard to the positions of the points in the map by defining a similar probability distribution over the points in the low-dimensional map. 
Although the original algorithm bases its similarity metric on the Euclidean distance between objects, this can be modified as necessary.

ML | T-distributed Stochastic Neighbor Embedding (t-SNE) Algorithm
https://www.geeksforgeeks.org/ml-t-distributed-stochastic-neighbor-embedding-t-sne-algorithm/

Read More