Which method changes the shape of the data (not just its scale) to make it easier to analyze or model, especially when data are highly skewed?

Which method changes the shape of the data (not just its scale) to make it easier to analyze or model, especially when data are highly skewed?

• A. Data Transformation
• B. Normalization (Min-Max)
• C. Standardization
• D. Principal Components Analysis (PCA)

Answer: (A)

Changing how the data look, not just how big or small it is, is what data transformation does. When data are highly skewed, applying a transformation—like a log, square root, or Box-Cox function—pulls in the long tails, compresses extreme values, and often makes the distribution resemble normal. That reshapes the data’s distribution and variance in a way that helps many modeling methods fit better, meet assumptions (like linearity and homoscedasticity), and produce more reliable inferences. Normalization (min–max) and standardization adjust the scale and variance of features but leave the shape of their distributions largely the same. PCA reshapes the data only in terms of representation: it rotates and projects data to capture maximum variance with fewer components, mainly for dimensionality reduction and decorrelation, not for correcting skewness. So the method described here is data transformation because its primary goal is to alter the data’s shape to aid analysis when skewness is present.

Changing how the data look, not just how big or small it is, is what data transformation does. When data are highly skewed, applying a transformation—like a log, square root, or Box-Cox function—pulls in the long tails, compresses extreme values, and often makes the distribution resemble normal. That reshapes the data’s distribution and variance in a way that helps many modeling methods fit better, meet assumptions (like linearity and homoscedasticity), and produce more reliable inferences.

Normalization (min–max) and standardization adjust the scale and variance of features but leave the shape of their distributions largely the same. PCA reshapes the data only in terms of representation: it rotates and projects data to capture maximum variance with fewer components, mainly for dimensionality reduction and decorrelation, not for correcting skewness. So the method described here is data transformation because its primary goal is to alter the data’s shape to aid analysis when skewness is present.