Офлайн-оценка дополняющих fashion-рекомендаций
Офлайн-оценка дополняющих fashion-рекомендаций
Ответить самому
Сначала сформулируйте ответ как на собеседовании, затем откройте разбор и оцените себя.
Короткий ответ
Evaluate retrieval with recall@K against outfit/compatibility labels, evaluate reranking with list quality and business proxies, and add human/style review for visual compatibility and diversity.
Полный разбор
Separate retrieval and reranking. For candidate generation, use outfit datasets, stylist labels, VLM-prelabeled data reviewed by humans, or historical co-engagement where appropriate. Measure recall@K, category coverage and how often a compatible item appears in the candidate set.
For reranking, pointwise metrics are not enough because the final list should look like a coherent outfit. Add list-level metrics such as category diversity, intra-list similarity, price/availability constraints, brand/category balance and business proxies such as expected conversion or revenue.
Finally, inspect examples with domain reviewers. Fashion compatibility has subjective and visual aspects, so offline numerical metrics should be paired with structured human review before online A/B testing.
Теория
Complementary recommendations need both relevance and list-level compatibility; retrieval recall alone does not prove the final outfit is useful.
Типичные ошибки
- Use only co-clicks and call them ground truth for style compatibility.
- Evaluate candidate generator and ranker with the same metric.
- Ignore category diversity in outfit recommendations.
- Skip human review for a visual-style product.
Как отвечать на собеседовании
- State which metric belongs to retrieval and which belongs to reranking.
- Mention stylist or human review as a calibration layer.