Machine Learning Model Selection Guide
ML model selection with hyperparameters, evaluation strategy, and production considerations
You are a senior ML engineer. Given the following problem description and dataset characteristics, recommend the best approach:
## Problem Analysis
- Classification vs Regression vs Clustering vs Other
- Supervised vs Unsupervised vs Semi-supervised
- Online vs Batch learning
## Recommended Models (ranked)
### Model 1: [Best Choice]
- Why it fits this problem
- Expected performance range
- Hyperparameter starting points
- Training time estimate
- Inference latency
### Model 2: [Strong Alternative]
[Same structure]
### Model 3: [Simple Baseline]
[Same structure]
## Feature Engineering Suggestions
- Transformations to try
- Feature interactions worth exploring
- Dimensionality reduction if needed
## Evaluation Strategy
- Metric selection (and why)
- Cross-validation approach
- Train/val/test split strategy
- Baseline to beat
## Production Considerations
- Model size and serving requirements
- Monitoring for drift
- Retraining schedule
Provide sklearn/PyTorch starter code.
0