Semantic search is an approach to finding content on the internet that anticipates the intention behind the user’s query. The goal of semantic search is to provide the end user with the most relevant search engine results possible.
In semantic search, the search engine’s programming identifies the keyword in a query but also tries to predict user intent when returning results. When predicting user intent, programming may factor in such things as previous searches, the user’s geographical location, trending topics, the relationship between words in the user’s query, the relative success of similar queries, ontology interrelationships and the type of device submitting the query.
Unlike Boolean search, which can only accommodate keywords and the operators AND, OR and NOT, semantic search allows users to use natural language when submitting queries. The programming uses fuzzy logic, predictive modeling and deep learning algorithms as well as text analytics, knowledge graphs and concept maps to provide the user with the order of links on a search engine results page (SERP).
The programming also gathers data about what links the end user clicks on, what links the user bounces back from quickly and metrics that indicate user engagement to improve future query results. The programming’s disambiguation capabilities can not only differentiate between two similar keywords, it can also recognize variations in spelling and verb tense.
Semantic search is often associated with Google RankBrain, the artificial intelligence (AI) component of Google's Hummingbird search algorithm. RankBrain uses machine learning to filter results and improve which results are positioned first in search engine results pages. RankBrain programming searches through data to find patterns and uses that data to improve the software's own understanding.