Before selecting features for a risk-analysis model, an ML engineer wants to estimate how much each training feature contributes to the target. Which SageMaker Data Wrangler approach is appropriate to obtain feature contribution or importance estimates?

SageMaker Data Wrangler's quick model visualization is specifically designed to provide insights into feature importance. It trains a simple model (like a LightGBM or XGBoost) on your dataset and then calculates and displays feature importance scores, directly estimating each feature's contribution to the target variable. Multicollinearity measurements and PCA (Principal Component Analysis) help understand relationships between features and reduce dimensionality, but they don't directly quantify a feature's contribution to the target variable in the same way feature importance scores do. The bias report identifies potential biases in the data, which is a different concern than feature contribution. Building a data-prep pipeline is a general step in data processing; manually assigning feature scores is subjective and not a data-driven method for estimating contribution.