Build a Machine Learning Model Pipeline

Build a machine learning pipeline for the following:

Problem type: [classification / regression / clustering / NLP / computer vision / recommendation / other]
Problem description: [describe what you are trying to predict or learn]
Data description: [features, target variable, size, format]
Success metric: [accuracy / F1 / RMSE / precision / recall / other]
Language: [Python]
Libraries preference: [scikit-learn / TensorFlow / PyTorch / XGBoost / HuggingFace / other]
Deployment target: [REST API / batch processing / edge / other]
Data privacy requirements: [any constraints on data usage]

Provide a complete ML pipeline:

DATA EXPLORATION:
- Exploratory data analysis code
- Key statistics and distributions to check
- Visualisations to create
- Data quality assessment

DATA PREPROCESSING:
- Missing value strategy (per feature type)
- Outlier detection and treatment
- Feature encoding (categorical variables)
- Feature scaling/normalisation
- Train/validation/test split strategy

FEATURE ENGINEERING:
- Feature creation ideas for this problem type
- Feature selection approach
- Feature importance analysis

MODEL SELECTION:
- Candidate models for this problem type
- Baseline model to beat
- Model comparison framework

MODEL TRAINING:
- Training code for recommended model(s)
- Cross-validation setup
- Hyperparameter tuning (GridSearch / RandomSearch / Bayesian)
- Training monitoring

MODEL EVALUATION:
- Evaluation metrics code
- Confusion matrix and error analysis
- Learning curves
- Bias and fairness checks

MODEL DEPLOYMENT:
- Model serialisation
- REST API wrapper code
- Input validation for inference
- Prediction endpoint

MONITORING AND MAINTENANCE:
- Data drift detection
- Model performance monitoring
- Retraining trigger criteria
- A/B testing new model versions