T-SNE
t-SNE is a nonlinear dimensionality-reduction technique that maps high-dimensional data to a lower-dimensional space while emphasizing local similarities or neighborhood relationships.
Data Without Labels
t-distributed stochastic neighbor embedding is a nonlinear dimensionality-reduction technique that maps high-dimensional observations to a low-dimensional space while preserving local similarities.
- How it is used
- Visualizes local similarities in high-dimensional data, including image measurements, through low-dimensional projections.
- How deep it goes
- Traces Gaussian and Cauchy similarities through KL-divergence minimization, then demonstrates fitting and plotting in Python.
- Examples
- Uses labeled Iris scatterplots and reshaped MNIST digit images to inspect two-dimensional projections.
- What you can do after
- Apply it via scikit-learn’s TSNE to create and interpret labeled two-dimensional plots, while checking perplexity choices.
Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI
t-SNE is a nonlinear dimensionality-reduction technique that maps high-dimensional data to a low-dimensional space while emphasizing local neighborhood structure.
- How it is used
- Supports exploratory visualization of complex tabular data by exposing patterns, clusters, and possible anomalies.
- How deep it goes
- Reduces numeric multivariate data to plotted dimensions, but stops short of stable interpretation because hyperparameters alter projections.
- Examples
- Uses numeric and ordinal dataset features, mean imputation, scikit-learn TSNE, and a two-dimensional scatter plot.
- What you can do after
- Apply it via scikit-learn's TSNE to create and plot a two-dimensional data projection.
Read how Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI explains t-SNE →
Also covered in
- Machine Learning Algorithms in Depth
- Financial AI in Practice
- Ensemble Methods for Machine Learning
- Fight Fraud with Machine LearningIn early access (MEAP)
- Hugging Face in Action
What to know first
- clusteringClustering is the process of grouping data items so that items within a group are more similar to one another than to items in other groups.
Related concepts
- dimensionality reductionkind of · Dimensionality reduction transforms data into a representation with fewer dimensions while attempting to retain relevant information or structure.
- Kullback–Leibler divergenceuses · Kullback–Leibler divergence is a nonnegative, generally asymmetric measure of the difference between probability distributions, defined as the expected log probability ratio under one distribution.
- embeddingproduces · An embedding is a numerical vector representation of an item in a feature space, encoding selected properties or relationships for computational tasks such as comparison or retrieval.
- principal component analysisalternative to · Principal component analysis is a linear statistical technique that represents data along orthogonal directions, called principal components, ordered by the variance they capture.
- UMAPalternative to · UMAP, or Uniform Manifold Approximation and Projection, is a nonlinear manifold-learning technique that embeds high-dimensional data in a lower-dimensional space while preserving important neighborhood and structural relationships.
Questions about t-SNE
What is t-SNE?
t-SNE is a nonlinear dimensionality-reduction technique that maps high-dimensional data to a lower-dimensional space while emphasizing local similarities or neighborhood relationships.
Which Manning books explain t-SNE?
Data Without Labels and Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI.
What should I know before learning t-SNE?
Manning books that teach t-SNE usually introduce clustering first.
What is the difference between t-SNE and principal component analysis?
t-SNE is a nonlinear dimensionality-reduction technique that maps high-dimensional data to a lower-dimensional space while emphasizing local similarities or neighborhood relationships. Principal component analysis is a linear statistical technique that represents data along orthogonal directions, called principal components, ordered by the variance they capture.
What is t-SNE a kind of?
T-SNE is a kind of dimensionality reduction. Dimensionality reduction transforms data into a representation with fewer dimensions while attempting to retain relevant information or structure.
What is t-SNE also called?
T-SNE is also called t-Distributed Stochastic Neighbor Embedding.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.

