A data scientist inspects a feature and finds the distribution mode < median < mean. Which transformation is most appropriate to transform this data so a linear regression model can be applied?

The distribution mode < median < mean indicates a right-skewed distribution. Linear regression models assume normally distributed residuals, and transforming skewed features can help achieve this. A logarithmic transformation (e.g., log(x), log(x+1)) is highly effective at reducing right skewness, pulling in the long tail of the distribution and making it more symmetrical, which is beneficial for linear models. Exponential transformation is used for left-skewed data or to increase variance. Polynomial transformation creates new features by raising existing features to a power, which doesn't directly address skewness in the same way. Sinusoidal transformation is for cyclical data and is irrelevant here.