Consider these two scenarios:
8 a.m.: In advance of a product launch, a pharmaceutical analyst is at his desk, searching for news and information related to recent drug patents, highlighting the active ingredient, the intended application/use, as well as who owns it and the date and country in which it is filed, and any other pertinent information related to its performance in the market or other relevant innovations.
7 p.m.: At home, the same analyst is looking for a used car, something along the lines of a Volvo wagon, with air bags, maximum 150,000 miles, automatic transmission, etc. The process of comparing offers, prices and options on a variety of websites, all expressed in a thousand different ways, takes hours, even days of searching to find valid options.
These are two common situations, which, although they exist in different spheres, deal with the same problem–how do we find the information we’re looking for–and the most relevant of that information, among all of the information available to us?
Documents, articles, web pages, emails and social media posts all contain information that could be useful in either of these cases, but think about the time and effort it takes to wade through all of that information in order to finally get to those things that truly match our request or need.
Text analytics is the process of selection and extraction of entities and relevant concepts in any kind of text. Traditional analysis methods are not effective enough to deal with the complexity of such texts–they are not ‘smart’, they don’t ‘understand’ language, and instead, are limited to managing lists already organized in a database.
Instead, an approach that simulates the human capacity to reason doesn’t just access the pre-defined criteria list, but by understanding their context, can also expand to include terms that, through meaning are relevant, even if you didn’t think to include them. This semantic approach proves accurate and reliable in finding the information you’re looking for–intercepting and connecting all of the information on pharmaceutical patents and used cars, and more.
These are great examples that illustrate the possibilities made available to us by semantics, and just another reason why we’re so optimistic about the future!