Databricks Machine Learning Associate Question 540
Select 3In which of the following scenarios is it appropriate to apply a log scale transformation to the data?
- A
When the data has a highly skewed distribution with large positive values.
- B
When the data contains negative values that need to be normalized.
- C
When the relationship between variables is multiplicative rather than additive.
- D
When the data contains categorical variables that need to be encoded.
- E
When the data spans several orders of magnitude.
Show answer and explanation
Correct answers: A, C, E
Explanation
Log scale transformation is commonly used in machine learning when dealing with numerical data that is highly skewed, spans large ranges, or has a multiplicative relationship between variables. It helps normalize the data, improve model interpretability, and stabilize variance.
- A. Correct.
Log scale transformation is useful for highly skewed data with large positive values, as it reduces skewness and makes the data more normally distributed.
- B. Incorrect.
Log scale transformation cannot be applied to negative values or zeros because the logarithm of a negative number or zero is undefined.
- C. Correct.
Log transformations are appropriate when the relationship between variables is multiplicative, as it converts the multiplicative relationship into an additive one.
- D. Incorrect.
Log scale transformation is not used for processing categorical data. Categorical variables require encoding techniques like one-hot encoding or label encoding.
- E. Correct.
Log scale transformation is effective when data spans several orders of magnitude, as it helps to compress the range and manage extreme values.