Millions of news articles are published every day, but only a small fraction of them are about the subjects and the industries that you care about, which could for instance be cryptocurrencies or the oil & gas industry, or perhaps a certain type of event like a natural disaster.
In order to identify these relevant articles, we use article categorization, which uses a natural language processing technique called classification.
Simply put, the way it works is: We identify the set of topics and categories we would like to be able to identify (this is often called a taxonomy), then we provide examples of articles pertaining to those categories which could be manually tagged, and finally we train a model that learns from the manually tagged articles to identify the topics in our taxonomy given a new unseen article that was just published.
Once article categorization is applied to millions of articles, as we do in our News API, you can accurately identify the documents you should care about.
AYLIEN’s neural network based categorization system is able to classify articles regardless of the language they’re written in into highly granular taxonomies which cover:
- 1500 Industries and sub-industries, covering major industrial classifications such as Banking & Finance, Energy & Utilities, and more.
- 2900 Subjects, covering broad and narrow subjects within Business & Commerce, Finance & Economics, or Health.
- 236 Trading Impact categories, covering potentially market moving events such as Bankruptcy & Insolvency, Credit Ratings, and Civil Unrest
- 71 Adverse News categories covering Adverse Events such as Natural Disasters, Layoffs, Product Recalls, and Strikes.
- Additionally we support standard IPTC and IAB category taxonomies.
To find out more about how categorization in AYLIEN News API helps you find the news that matters to you, sign up for a free trial here.