Dimension Reduction

Correlation analysis is used for

• Detecting relation between two variables

• Check for Multicollinearity

• Detect duplication of variables

• All of the above

Which plot can be used to show correlations

• Box plot

• Bar Plot

• Side by side correlation plot

• Heatmaps

PCA

• Reduces number of observations

• Reduces number of variables

• Detects pattern between variables

• Detects the components in a variable

Which value should you look at to determine how many PCs to use?

• Mean

• Standard deviation

• Proportion of Variance

• Cumulative Proportion

To remove scale effect and reduce the dominant effect of any variable we use

• Normalization

• PCA

• Linear Regression

• Variance Reducer