# Değerlendirme

Canonical: https://manyspace.aporia.systems/tr/subjects/ai/evaluation-115
Updated: 2026-10-13

Değerlendirme: yaklaşık 16 saat. Her konu önce neye ihtiyaç olduğunu ve neyi açtığını listeler.

## Discipline

- [Offline Evaluation](https://manyspace.aporia.systems/tr/topics/offline-evaluation-aiml-eval-001)
- [Statistical Significance and Confidence Intervals for Evaluation](https://manyspace.aporia.systems/tr/topics/statistical-significance-and-confidence-intervals-for-evaluation-aiml-eval-012)
- [Online Experiments and A/B Testing](https://manyspace.aporia.systems/tr/topics/online-experiments-and-a-b-testing-aiml-eval-015)
- [Baselines and Ablations](https://manyspace.aporia.systems/tr/topics/baselines-and-ablations-aiml-eval-016)

## Supervised Metrics

- [Regression Metrics](https://manyspace.aporia.systems/tr/topics/regression-metrics-aiml-eval-002)
- [Classification Metrics](https://manyspace.aporia.systems/tr/topics/classification-metrics-aiml-eval-003)
- [Confusion Matrices](https://manyspace.aporia.systems/tr/topics/confusion-matrices-aiml-eval-004)
- [ROC and Precision-Recall Curves](https://manyspace.aporia.systems/tr/topics/roc-and-precision-recall-curves-aiml-eval-010)
- [Ranking Metrics](https://manyspace.aporia.systems/tr/topics/ranking-metrics-aiml-eval-011)

## Generation Eval

- [Generation Metrics](https://manyspace.aporia.systems/tr/topics/generation-metrics-aiml-eval-005)
- [LLM Evaluation](https://manyspace.aporia.systems/tr/topics/llm-evaluation-aiml-eval-006)
- [Human Evaluation](https://manyspace.aporia.systems/tr/topics/human-evaluation-aiml-eval-007)
- [Benchmark Contamination and Validity](https://manyspace.aporia.systems/tr/topics/benchmark-contamination-and-validity-aiml-eval-013)
- [LLM-as-a-Judge](https://manyspace.aporia.systems/tr/topics/llm-as-a-judge-aiml-eval-014)

## Trust

- [Robustness Evaluation](https://manyspace.aporia.systems/tr/topics/robustness-evaluation-aiml-eval-008)
- [Error Analysis and Slicing](https://manyspace.aporia.systems/tr/topics/error-analysis-and-slicing-aiml-eval-009)
