Suppose you have the search query “Practical Semantic Search Books”. (It's gratifying to see my Google Semantic search book dominates the search results). In order for Google Search to find the answer to that search query across the web it would have to create, internally, a binary classifier that looks at every web page across the web and uses them to create two data sets.
One marked “Relevant” and the other marked “Non-relevant”. The web however is a massively big place. Such classification would take a long time even at current search engine computational speeds.
To do their job search engines take shortcuts. They use a number of relational analysis and association factors to help find the most promising candidate for each of the two data sets.
It then becomes easier to expand the use of the same criteria to pick out the most promising candidate from the relevant data set.
Social media sharing and engagement by way of comments, mention and citations may not be a ranking factor but it is vital to your website being found through search.
Get smart: SEO Help: 20 Semantic Search Steps that Will Help Your Business Grow is a practical step-by-step guide to applying semantic search principles to your business.