Machine Learning Problem Bible (MLPB) is a collection of machine learning problems and solutions. It attempts to solve a few problems.
One of the best ways to start solving a problem is to analyze the solution to a similar problem. MLPB is an organized collection of machine learning problems and solutions, each with tags like “regression”, “natural-language-processing”, “hierarchical-data”, “random-forest”, etc. making it easy to identify problems and solutions related to the one at hand.
MLPB contains example solutions using the same model (e.g. gradient boosting) from different implementations (e.g. Python’s scikit-learn package, or XGBoost). This makes it easy to compare benefits and drawbacks of different programming languages and packages - not just different models.
All the examples in MLPB contain small datasets (thousands of rows or less), keeping the focus on understanding algorithms. This makes it easy to check intuitions about a model’s behavior and performance.