Skip to main content

Run

Run lets you run dimension reduction to transform your data from a high-dimensional space into a low-dimensional space while retaining meaningful properties of the original data.

Format​

Run uses a single format: Run dimensional reduction with <method> extracting <number> components (excluding | using) <columns> setting identifier as <target>, setting miminum distance as <distance>.

Parameters​

Run uses the following parameters:

  • method (required). The dimensional reduction method to use, selectable between PCA and UMAP.
  • number (required). The number of components to extract.
  • columns (required). The columns to exclude or include in the extraction.
  • target (optional). The column to set the identifier as. This column persists in the resulting dataset.
  • distance (optional). The value to set the minimum distance to. This parameter aids in balancing the preservation of local and global structures in the data.

Output​

If the dimensional reduction is successful, a new dataset appears in the Data tab. A corresponding chart also appears in the Chart tab that explains the variance ratio. Otherwise, an error message is shown in the conversation history.

Examples​

To dimensionally reduce a dataset called Heart_Health, enter Run dimensional reducation with PCA extracting 3 components using BMI, Sex, AgeCategory setting idetifier as HeartDisease

Feedback