Tools:Google use of semantic search features

From MarineLives
Jump to: navigation, search

Useful blog article: Paul Grabowski, 'Understanding Semantic Search - Introduction for Beginners', Feb 18th 2015

The article lists other resources on semantic search, including:


The Structured Search Engine
- Discusses Google acquisition of "Freebase" (now closed down as an independent entity)[1]
- To create a new Freebase entity extract tabular and attribute data in a web page
- Using Open-Domain Fact Extraction; rank extracted attributes with confidence values
- Query parsing, e.g. "when was martin luther king jr born"
- Parser identifies entities (thing being asked about, e.g. "martin luther king jr") and attributes (e.g. "born"); synonyms for specific entities; question forms (what "value" does the wusetion form found tend to deliver (here it is "date")
- Understanding content: Sentiment Analysis (positive vs. negative; happy vs sad) using Natural Language Processing. Example of summarising restaurant reviews.



People also ask


Google search: People also ask dropdown box

Google searches using common search terms generate first page boxes containing dropdown menus of frequent searches using that common search term.

They also include results from Google images and selected news featuring that search term.

For example, Google search for the term "Basketball" will yield a box containing dropdown questions related to basketball, which are frequently asked by searchers interested in basketball.



Searches related to specific search term or terms


Google search: Searches related to specific search term or terms

Google searches using a single or multiple search terms will generate a table of "Searches related to..." at the bottom of the first page of search results.

For example, Google search for the term "cheapest apple" will yield a box with eight further suggested searches using those terms

Specific Google services


  1. Wikipedia article: Freebase