Session Ready
Exercise

Singular value decomposition

In the last exercise, you saw how useful PCA could be in reducing the dimensions of a dataset when you're given a question about high dimensionality in a machine learning interview.

In this exercise, you'll practice SVD on the diabetes. This particular transformer can work with sparse matrices efficiently, as opposed to PCA, and performs linear dimensionality reductions by way of truncated singular value decomposition.

Recall that singular value decomposition takes the original data matrix, decomposes it into three matrices and uses them to calculate and return singular values.

Same place in the pipeline with a different technique: Machine learning pipeline

Instructions 1/4
undefined XP
  • 1
  • 2
  • 3
  • 4
  • Import the relevant module to perform SVD.
  • Create a feature matrix X and target array y with progression from the diabetes dataset.