Selections
Event selection algorithms for cut-based physics analyses.
Overview
The selections module contains selection implementations for various studies:
analysis: Main analysis selections with regions
example: Example selections for tutorials
mc16: MC16 truth studies
mc20: MC20 truth studies
neutrino: Neutrino reconstruction studies
performance: Performance evaluation selections
Selection Organization
Selections are organized by:
Physics Analysis: Different final states (4-tops, ttbar, etc.)
Selection Regions: Signal, control, validation regions
Study Type: Truth matching, kinematics, reconstruction
MC16 Selections
Truth-level studies on MC16 samples:
childrenkinematics: Top decay product kinematics
decaymodes: Top quark decay mode classification
met: Missing transverse energy studies
parton: Parton-level analysis
topjets: Top-tagged jet studies
topkinematics: Top quark kinematic distributions
topmatching: Truth matching algorithms
toptruthjets: Truth jet associations
zprime: Z’ resonance searches
MC20 Selections
Studies on MC20 simulation samples:
matching: Object matching algorithms
topkinematics: Top kinematic distributions
topmatching: Truth matching performance
zprime: Heavy resonance searches
Neutrino Selections
Neutrino reconstruction studies:
combinatorial: Combinatorial reconstruction methods
validation: Validation of reconstruction algorithms
Performance Selections
Framework performance studies:
topefficiency: Top reconstruction efficiency
Selection Template Usage
All selections inherit from selection_template and implement:
Selection(): Define event selection logic
InitHistograms(): Book histograms
ApplySelection(): Apply cuts and fill histograms
Finalize(): Produce final plots and statistics
Cutflow Tracking
Selections automatically track:
Number of events at each cut
Efficiency of each cut
Cumulative selection efficiency
Statistical uncertainties
This enables:
Optimization of cut ordering
Understanding selection performance
Debugging selection logic
Reporting results
Output
Selections produce:
Histograms: Kinematic distributions
Tables: Cutflow statistics
ROOT Files: For further analysis
Plots: Publication-quality figures
Integration with Analysis
Selections are executed via the Analysis framework:
Events are loaded from ROOT files
Events are passed to Selection instances
Selections apply cuts and fill histograms
Results are aggregated across all events
Final outputs are produced
This allows:
Processing large datasets efficiently
Running multiple selections in parallel
Caching intermediate results
Modular analysis workflows