Chapter 8. Providing relevance feedback
This chapter covers
- The ongoing, multifaceted conversation between user and search
- Methods, besides relevance ranking, for getting users to relevant content, including
- Guiding users toward better search queries
- Correcting users’ searches through spell-checking
- Highlighting why documents are relevant to user searches
- Explaining to users how their search is interpreted
- Allowing users to filter out irrelevant content from the result set
To this point, we’ve discussed how to deeply manipulate relevance ranking. In this chapter, you’ll see that relevance ranking isn’t the only way to guide users to relevant content. Search enables a multifaceted conversation between the user and the search engine. Because the relevance portion of this conversation is never perfect, your users will always thank you for additional guidance toward relevant content. In this chapter, you’ll steer the many layers of the search conversation beyond relevance ranking by building features that do the following:
- Explain to users how their query is being interpreted
- Correct mistakes such as typos and misspellings
- Suggest other searches that will provide better results
- Convey an understanding of how documents are distributed in the corpus
- Help users understand why a particular document is a match
- Help users efficiently understand the result set