- Precision (P)
- Average Precision (AP)
- Cumulative Gain (CG)
- Discount Cumulative Gain (DCG)
- normalized DCG (nDCG)
The conclusion explains it very well, so I am going to copy and paste it here:
There isn’t a single metric that will give you a perfect summary value for your search’s performance, but a combination of metrics can get you close. How search results are displayed (large-grid vs vertical-list) should influence which metric you choose, because accounting for position/rank makes less sense when results are shown in bulk and meant to be perused.
Also remember that these metrics will be used primarily as you iterate on your search engine configuration, comparing different system configurations. So when you are testing if a new system configuration is really an improvement, you will be looking at the change in a metric for the same query across the two configurations.
Read the whole article here: https://opensourceconnections.com/blog/2020/02/28/choosing-your-search-relevance-metric/