A heuristic is a rule-based method that produces a practical solution without learning a model from data. In an ML project, test whether a simple heuristic is sufficient before committing to the cost and complexity of machine learning. section 1.1.1
Simple heuristics can be efficient, but complex, high-dimensional data with intricate and nonlinear patterns may require an ML model for accurate analysis or prediction. Treat the choice as an explicit problem-formulation decision rather than an automatic appeal to ML. section 1.1.1
Use a heuristic when:
- A straightforward rule appears likely to solve the operational problem efficiently. section 1.1.1
- Avoid replacing the rule with ML merely because ML is available; escalate when data complexity demands learned patterns. section 1.1.1
