Glossary term
Glossary term
Foundations
Decreasing the number of dimensions used to represent a particular feature in a feature vector, typically by converting to an embedding vector.
Created for this library
A marketing analytics team uses dimension reduction to compress a 200-column customer profile into 20 dimensions for visualization and clustering.
A genomics team uses dimension reduction on gene expression profiles to visualize cell types in two dimensions for biologists.
A retail analytics team uses dimension reduction on product-purchase histories to build a compact representation for collaborative filtering.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License