A correlation matrix is a square table listing the correlation coefficients between pairs of attributes.

• A. False
• B. Not necessarily
• C. True
• D. Only for numeric attributes

Answer: (C)

A correlation matrix is a compact way to show how numeric attributes relate to each other. For a dataset with m numeric attributes, you compute a correlation coefficient for every pair and place it in an m-by-m table. The result is square because every pair of attributes gets a value, and the matrix is symmetric since the correlation of A with B is the same as B with A. The diagonal entries are 1, since each attribute is perfectly correlated with itself. This format lets you quickly spot strong positive or negative relationships and compare them across all pairs.

Note that standard correlation measures require numeric data, so for non-numeric attributes you’d either restrict to the numeric subset or use different measures. That's why the statement is true in the usual context: a correlation matrix lists the pairwise coefficients between numeric attributes.