moorcheh/mair-ndcg10-results-all-providers

该数据集包含全面的**NDCG@10（归一化折损累积增益）**准确性结果，涵盖了各种向量数据库提供商和检索配置。基准测试比较了在多个专业领域（法律、金融、医疗和API文档）中使用**量化**和**浮点**向量的性能。结果比较了Moorcheh与行业标准（如Elasticsearch、Pinecone、PGVector、Qdrant等）的检索准确性。数据集结构包括数据集名称、类别、数据集大小以及各提供商在量化和浮点向量下的NDCG@10分数。性能总结显示，Moorcheh在法律和金融领域通常优于或匹配标准PGVector和Qdrant实现，同时量化对精度的影响较小。

This dataset contains comprehensive **NDCG@10 (Normalized Discounted Cumulative Gain)** accuracy results across various vector database providers and retrieval configurations. The benchmarks compare performance using both **Quantized** and **Floating-Point** vectors across multiple specialized domains (Legal, Financial, Medical, and API Documentation). The results compare the retrieval accuracy of **Moorcheh** against industry standards like Elasticsearch, Pinecone, PGVector, and Qdrant. The dataset structure includes dataset names, categories, dataset sizes, and NDCG@10 scores for each provider under both quantized and floating-point vectors. The performance summary highlights that Moorcheh often outperforms or matches standard PGVector and Qdrant implementations in legal and financial domains, with minimal accuracy drop due to quantization.