InhaltsverzeichnisSeite 1
Plan
Evaluating search engines
Measures for a search engine
Seite 5
Measuring user happiness
Seite 7
Happiness: elusive to measure
Evaluating an IR system
Standard relevance benchmarks
Precision and Recall
Accuracy
Why not just use accuracy?
Precision/Recall
A combined measure: F
F1 and other averages
Ranked results
Precision-recall curves
Interpolated precision
Evaluation
Difficulties in using precision/recall
|