t-SNE is a manifold learning technique, which learns low dimensional embeddings for high dimensional data. It is most often used for visualization purposes because it exploits the local relationships between datapoints and can subsequently capture nonlinear structures in the data. Unlike other dimension reduction techniques like PCA, a learned t-SNE model can't be applied to new data. The t-SNE algorithm can be roughly summarized as two steps:
- Create a probability distribution capturing the relationships between points in the high dimensional space
- Find a low dimensional space that resembles the probability dimension as well as possible
Disclaimer:
Depending on the size of the input table, the computation of t-SNE can be very expensive both in terms of runtime as well as memory. If you experience problems with memory, try to reduce the size of your data by e.g. using the Row Sampling node. If you have very high-dimensional data, it is also advisable to first reduce the number of dimensions to around 50 using e.g. a PCA.Options
Popular Successors
- Scatter Plot (JavaScript)21 %
- Color Manager17 %
- 2D/3D Scatterplot6 %
- k-Means5 %
- 3D Scatter Plot (Plotly)4 %
- Scatter Plot (Plotly)4 %
- Column Filter3 %
- Scatter Plot (JFreeChart)2 %
- Joiner2 %
Column Rename2 %- Rank Correlation2 %
- RowID2 %
- CSV Writer2 %
Joiner2 %- t-SNE (L. Jonsson)1 %
- Excel Writer1 %
- Column Appender1 %
- Normalizer1 %
- Scatter Plot (legacy)1 %
Column Appender< 1 %- String Manipulation< 1 %
- Category to Number< 1 %
- X-Partitioner< 1 %
- Row Splitter< 1 %
- Sorter< 1 %
- Timer Info< 1 %
- Number to String< 1 %
- Shape Manager< 1 %
- Table to Spark< 1 %
- Distance Matrix Calculate< 1 %
- 2D Density Plot (Plotly)< 1 %
- Decision Tree Learner< 1 %
- Auto-Binner< 1 %
- Top k Row Filter< 1 %
- Numeric Distances< 1 %
- GroupBy< 1 %
- Correlation Filter< 1 %
- Numeric Outliers< 1 %
- Kolmogorov-Smirnov Test< 1 %
- Contour Plot (Plotly)< 1 %
- Histogram Chart (JFreeChart)< 1 %
- Tableau Writer< 1 %
- R Snippet< 1 %
- Graph Density Initializer< 1 %
Table Row to Variable< 1 %CSV Writer< 1 %Counting Loop Start< 1 %- Silhouette Coefficient< 1 %
- Fuzzy c-Means< 1 %
- Hierarchical Clustering< 1 %
- Hierarchical Cluster Assigner< 1 %
Cell Replacer< 1 %- Column Aggregator< 1 %
- Column Merger< 1 %
- Column Resorter< 1 %
- Nominal Value Row Filter< 1 %
- Row Filter< 1 %
- Reference Row Splitter< 1 %
- Partitioning< 1 %
- Row Sampling< 1 %
- String Manipulation (Variable)< 1 %
- String Replacer< 1 %
- Rule-based Row Filter< 1 %
- Independent groups t-test< 1 %
Color Appender< 1 %- Size Manager< 1 %
- Math Formula< 1 %
Excel Writer (XLS)< 1 %- Table View (JavaScript)< 1 %
Python Script (1⇒1)< 1 %- R Snippet< 1 %
- R View (Table)< 1 %
Table Writer< 1 %- Logistic Regression Predictor< 1 %
- Column Combiner< 1 %
- Domain Calculator< 1 %
- Duplicate Row Filter< 1 %
- HiLite Row Splitter< 1 %
- Rule Engine< 1 %
- Scatter Matrix (legacy)< 1 %
- Gaussian Distributed Assigner< 1 %
- k-Medoids< 1 %
- Autocomplete Text Widget< 1 %
- Bubble Chart (JFreeChart)< 1 %
- Rule Viewer (local) (Legacy)< 1 %
- Nominal Row Filter Configuration< 1 %
- Python Learner (legacy)< 1 %
- Python View (legacy)< 1 %
- Data to Report (BIRT)< 1 %
- RDKit From Molecule< 1 %
Views
- This node has no views
Workflows
(3D) Fragment library design workflowKNIME Hub- 01_BERT_Sentiment_AnalysisKNIME Hub
- 02_PCA_t-SNEKNIME Hub
- 02_Techniques_for_Dimensionality_ReductionKNIME Hub
- 3.3_ChemicalSpaceKNIME Hub
- Show all 39 workflows
Links
Developers
You want to see the source code for this node? Click the following button and we’ll use our super-powers to find it for you.