Chapter 1. The search relevance problem
Figure 1.1. The relevance engineer works with the search engine and back-end technologies to express business-ranking logic. They collaborate on relevance closely with a cross-functional team and are informed heavily by user metrics.
Figure 1.2. MeSH categorization of “Myocardial Infarction” (left) along with several MeSH topics closely related to “Myocardial Infarction”
Figure 1.3. Example of making a relevance judgment for the query “Rambo” in Quepid, a judgment list management application
Figure 1.4. Relevance engineers select, enrich, or create important features from back-end systems and express ranking signals in terms of those features.
Figure 1.5. Forms of search-relevance feedback
Chapter 2. Search—under the hood
Figure 2.1. A simple model of a search engine based on possible interactions
Figure 2.2. Typical search user interface and response page
Figure 2.3. “Barcelona Beaches” article indexed and analyzed (only title-field analysis shown)
Figure 2.4. The inverted index data structure used by search engines closely resembles the index that you can find in the back of a textbook.
Figure 2.5. The full search ETL pipeline: extraction, enrichment, analysis, and indexing
Figure 2.6. Analysis—character filtering
Figure 2.7. Analysis—tokenizing
Figure 2.8. Analysis—token filtering
Figure 2.9. Search facets as presented on a Zappos search results page
Figure 2.10. Example document-scoring function (simplified)