Doug Turnbull is a search relevance consultant at OpenSource Connections where he frequently speaks and blogs. Using solutions like Elasticsearch, Doug builds relevant, semantically enriched search experiences for clients across multiple domains. John Berryman is a data scientist at EventBrite where he specializes in recommendations and search. He is interested in the potential of integrating semantic understanding into search and discovery applications.
It's easy to hit a wall when improving the quality of search results. Search relevance ranking can feel very mystical. Leveraging the features of the search engine to return relevant search results is a challenging art. Improving relevance means crafting relevance ranking for a particular kind of content, for a specific user audience with their own expectations, expertise, and vernacular. For this reason, relevance work is unlike any other technical task. Despite the difficulty, the practice of improving relevance is extremely important. Even search developers like us can lose track of how often we use search -- how quickly it has become the *entire* user experience. As users, we don't organize and browse anymore; we search. We look up contacts and friends, find lawnmower parts on craigslist. We search for music, emails, or movies to watch. Doctors search for the latest techniques at the bedsides of sick patients. Patent examiners search existing patents to identify prior art. Search is eating the world -- sinking its fangs into every application. Given search's increasing ubiquity, improving the quality of answers given by the search engine is a keenly felt need.
To address this need, we're writing Taming Search to teach the practice of search relevance. Search is creeping into so many applications and workflows. The needs of each application -- the very definition of relevance -- changes per user experience, user audience, and content. Yet the practice of improving search relevance has common ground and Taming Search teaches it to you!
Taming Search bridges two areas of wisdom. On the one hand, there's academic works like Introduction to Information Retrieval that teaches you the computer science, heuristics, and natural language processing behind building a generally relevant search engine. On the other hand, there are practical books like Lucene in Action or Elasticsearch: The Definitive Guide that provide an overview of Lucene technologies and their features, but don't delve search relevance in any great depth.
To improve the search results of everyday applications, we need to bridge the two areas of wisdom. How can you apply the lessons of Information Retrieval to today's search engines? What practical search lessons aren't captured in academic literature that nobody tells you? What practices do the experts use when improving relevance? How can external, enriching resources like ontologies or machine learning technologies be brought to bear on the problem? Taming Search answers these questions and more!
So if you struggle with poor search results, think relevance scoring is a mystical black box, or just want to gain a deeper appreciation of the seeming magic in the search engine, then Taming Search is the book you've been searching for! Check it out, and please be part of the conversation by participating in Taming Search's forum.
For a limited time, Manning has extended an early-access discount code to Elastic blog readers! Receive 50% off with code “mlturnbull"! So buy now!