A priority score measures a feature’s value relative to the effort required. Here, it is calculated as value points divided by effort points and used to rank competing features. section 8.2.2 section 8.2.3
Use priority scores to support constrained feature selection:
- Represent each feature with value points, effort points, first-year cost, total cost, and its strategic epic. section 8.2.2
- Maximize the combined scores of selected features while respecting budget limits and strategic balance. section 8.1.3 section 8.1.4
- Treat selection variables as binary: 1 selects a feature and 0 leaves it unselected. section 8.2.2
- Negate scores when using a minimization-oriented solver so minimizing the transformed objective maximizes the original priority. section 8.2.3
Automated Data Cleaning Pipeline
With 16 value points and 13 effort points, the feature receives a priority score of 16 ÷ 13 = 1.23. A score of 1.45 for Feature Engineering Automation indicates a higher value-to-effort ratio, although cost and epic constraints still determine whether either feature is selected. section 8.1.4 section 8.2.2
>>> c = [-f.value_points / f.effort_points for f in features]
Inspect the compact Python expression that computes the negated value-to-effort coefficient for every feature. section 8.2.3
