Improve ML basics tutorial notebook

[duhmariya]

Improved ML Basics Tutorial Notebook

Changes:

• Fixed data leakage: StandardScaler is now applied correctly within train/test split (not before cross_val_score)
• Added regression section (Linear Regression on Diabetes dataset with MSE and R² evaluation)
• Added unsupervised learning section (K-Means clustering)
• Added overfitting/underfitting demonstration (KNN with varying k from 1 to 30)
• Improved visualizations with labeled axes, titles, and confusion matrix
• Added variance explained analysis for cross-validation

Motivation:

The previous notebook had a scaling issue where X_train was standard-scaled before being passed to cross_val_score, causing data leakage. This version restructures the pipeline to follow ML best practices while covering classification, regression, and clustering.