Which technique is most appropriate to handle skewed numerical data in a dataset?

Log transformation is effective for handling skewed numerical data, as it compresses large values and stretches smaller ones, reducing skewness and stabilizing variance. It ensures the data conforms to normality, which many statistical methods and machine learning algorithms assume. Transformations like square root or cube root are alternative options for specific distributions. Why Other Options Are Wrong : B) One-hot encoding is for categorical data, not numerical skewness. C) Normalization rescales data but doesn’t address skewness effectively. D) Replacing outliers addresses extreme values but doesn’t fix overall skewness. E) Removing skewed columns discards valuable information and reduces dataset size.