Related Data Science Links
Learn Dim Reduction Data Science Tutorial, validate concepts with Dim Reduction Data Science MCQ Questions, and prepare interviews through Dim Reduction Data Science Interview Questions and Answers.
Dimensionality Reduction Q&A
1What is dimensionality reduction?
Answer: Reducing feature count while preserving useful information.
2Why use it?
Answer: Faster training, less overfitting, better visualization.
3PCA in one sentence?
Answer: PCA projects data onto orthogonal directions of max variance.
4PCA needs scaling?
Answer: Usually yes, to prevent large-scale features dominating components.
5Explained variance ratio?
Answer: Fraction of total variance captured by each component.
6Feature selection vs extraction?
Answer: Selection keeps original features; extraction creates new components.
7What is t-SNE used for?
Answer: Nonlinear 2D/3D visualization of high-dimensional data neighborhoods.
8What is UMAP?
Answer: Fast manifold learning method for structure-preserving embeddings.
9Can PCA improve model performance?
Answer: Sometimes, by denoising and reducing multicollinearity.
10Risk of too much reduction?
Answer: Information loss and poorer downstream performance.
11How choose number of components?
Answer: Cumulative explained variance and validation performance.
12One-line summary?
Answer: Dimensionality reduction simplifies data while retaining core signal.