HARRIS (Hybrid-decision trees)

[BrianSchiller]

No description provided.

[codecov]

Codecov Report

❌ Patch coverage is 97.26562% with 7 lines in your changes missing coverage. Please review.
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
asf/selectors/hybrid_decision_tree.py	97.56%	4 Missing ⚠️
asf/selectors/parallel_portfolio_selector.py	96.66%	3 Missing ⚠️

@@            Coverage Diff             @@
##             main     #137      +/-   ##
==========================================
+ Coverage   80.63%   81.53%   +0.89%     
==========================================
  Files          72       74       +2     
  Lines        4476     4732     +256     
==========================================
+ Hits         3609     3858     +249     
- Misses        867      874       +7     

Flag	Coverage Δ
full	81.53% <97.26%> (+0.89%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
asf/selectors/__init__.py	100.00% <100.00%> (ø)
asf/selectors/parallel_portfolio_selector.py	96.66% <96.66%> (ø)
asf/selectors/hybrid_decision_tree.py	97.56% <97.56%> (ø)

[Copilot]

Pull request overview

This PR introduces two new algorithm selectors for the ASF framework: HARRIS (Hybrid Ranking and Regression Forests) and APPS (Automatic Parallel Portfolio Selector). These selectors represent different approaches to algorithm selection - HARRIS uses a hybrid loss function combining regression and ranking objectives in decision tree ensembles, while APPS constructs parallel portfolios based on probabilistic performance predictions.

Key Changes:

• Implementation of HARRIS selector using hybrid decision trees with configurable lambda parameter to balance regression (MSE) and ranking (Spearman correlation) objectives
• Implementation of APPS selector that predicts algorithm portfolios for parallel execution based on winning probability thresholds
• Comprehensive example scripts demonstrating both selectors with synthetic data and performance evaluation

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 12 comments.

Show a summary per file

File	Description
asf/selectors/hybrid_decision_tree.py	Core implementation of HARRIS selector and HybridDecisionTree class with hybrid loss function
asf/selectors/parallel_portfolio_selector.py	Implementation of APPS selector with uncertainty-aware portfolio selection
asf/selectors/__init__.py	Exports both new selectors (APPS and HARRIS) to the public API
test/selectors/test_selectors.py	Basic test coverage for HARRIS and comprehensive tests for APPS including threshold behavior
examples/selection/18_harris_selector.py	Demonstration of HARRIS with different lambda values and performance evaluation
examples/selection/17_parallel_portfolio.py	Demonstration of APPS with various p_intersection thresholds and portfolio size analysis

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.