Chapter 9. Designing a relevance-focused search application
This chapter covers
- Gathering information before building a new search application
- Designing and implementing a complete search application
- Designing a query as a composite of subqueries
- Balancing query parameters
- Deploying, monitoring, and improving search
- Knowing when further relevance tuning is no longer advantageous
In the previous chapters, we laid out all the ingredients for good search:
- Extracting features from the text of the documents through proper tokenization
- Defining important signals and creating search fields to represent them
- Crafting queries that take into account both user needs and business requirements
- Providing feedback to users in order to guide them to more-relevant results
Now, this chapter is the rest of the recipe—the set of instructions that organizes these ingredients and lays out the methodology for building a relevance-focused search application.