subtopic

A practical algorithm decision guide

Fill the 'Algorithm Cards' sheet for both projects: which models to try first and why. Concepts: Data size, feature types, linearity, interpretability, cost, latency, deployment Project A: Small (1,000 rows), mixed categorical, regulated → logistic regression and a tuned gradient boosting are the main candidates. Project B: Large (~35k rows), strong seasonality, non-linear temperature effect → gradient boosting with lag/calendar features; linear model as baseline.