ElasticsearchBM25 - Part 3: Considerations for Picking b

Posted 九师兄

tags:

篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了ElasticsearchBM25 - Part 3: Considerations for Picking b相关的知识,希望对你有一定的参考价值。

在这里插入图片描述

This is the final post in the three-part Practical BM25 series about similarity ranking (relevancy). If you\'re just joining, check out Part 1: How Shards Affect Relevance Scoring in Elasticsearch and Part 2: The BM25 Algorithm and its Variables.

Picking b and k1

It’s worth noting that picking b and k1 are generally not the first thing to do when your users aren’t finding documents quickly. The default values of b = 0.75 and k1 = 1.2 work pretty well for most corpuses, so you’re likely fine with the defaults. More likely

以上是关于ElasticsearchBM25 - Part 3: Considerations for Picking b的主要内容,如果未能解决你的问题,请参考以下文章

2018.2.25-26 algo part3 greedy algorithm

[Stanford Algorithms: Design and Analysis, Part 2] c25 HUFFMAN CODES

R语言观察日志(part25)--将某列设置为行名

2018-06-25-Python全栈开发day21-part2-模块介绍

Redshift - Split_Part 和最小值

[原创]java WEB学习笔记25:MVC案例完整实践(part 6)---新增操作的设计与实现