Explanation

K-means clustering is the appropriate algorithm for this scenario because:

• The bank wants to group customers based on demographic features (age, marriage status, salary)
• This is an unsupervised learning task - there are no predefined labels for the groups
• K-means is specifically designed for customer segmentation and grouping similar data points

Why other options are incorrect:

• A: Natural language processing is for text analysis, not numerical clustering
• B: Logistic regression is for classification with known labels, not grouping
• C: PCA is for dimensionality reduction, not customer segmentation

K-means clustering is widely used in banking for customer segmentation based on demographic and financial characteristics.