Relevance Tuning: Search Weights
Search’s relevance settings can help you deliver the most relevant content to your users by letting you tailor how search queries match your results.
Relevance score
When you search, the system will calculate how well a document matches your query and give that document a relevance score (called a document score). This is used to deliver the most relevant results first. Relevance tuning is about tweaking this score calculation so that the results are ordered the way you want.
Weights
Weighting allows you to prioritise a field in your documents (see the Engines and Schema guide for more information on fields). An example of this could be two products that reference each other:
Product One
Title: Cat Ear Headphones
Subtitle: Cute animal-ear headphones
Description: High quality over-ear headphones with a feline quirk.
Product Two
Title: Halloween Headphones
Subtitle: Witches' cat headphones
Description: Like the Cat Ear Headphones but with a spooky twist ready for your Halloween party
In this example the word cat appears once in Product One but twice in Product Two. If you were to search for cat then, by default, product two would appear first as the most relevant because it contains more keyword matches. To avoid this you can add a weight to the Title field telling the search that matches in the Title field contribute more to the document score.
You can add weights to your searches via code using the Discoverer module from SDK. Larger weight values will mean the weighted field contributes more to the relevance score for a document.