Traditional Internet filtering methods depend on manually compiled blocking lists, individual ratings or online applied heuristics algorithms. These methods are, for the most part, inadequate, cannot keep up with the growth of the Internet or result in high numbers of false positives. As a consequence, inappropriate content is often allowed through the filter while acceptable content is blocked. Cobion instead uses a new approach to Internet filtering. The Cobion Content Filtering process, methodically and automatically scans the complete Internet and categorizes each website by its content using a proven combination of intelligent text classification and superior image recognition methods.
The scanning of the Internet is performed by Cobions Supercrawler, which inspects millions of new and updated websites every day. The categorization of all websites is done automatically using advanced technologies and is powered by a super computing infrastructure which provides the computing power that is necessary for this process. The result is a fresh and daily updated database of the Internet.
Cobion analyzes and independently categorizes Internet content into 58 categories. Currently, Cobion provides customers with a URL Database that contains more than 13 million categorized web page entries. This knowledge is based on the inspection and categorization of more than 1.8 billion web pages and images from the Internet. Cobion started this process in 1999 and since then has improved the quality of the content filtering process, expanded computing infrastructure and implemented new technologies to become more accurate and more up-to-date every day.
The service that Cobion provides to its customer is available under the name COFS (Content Filtering Service). This service contains the fresh and daily updated Cobion URL Database with 58 categories and is provided along with content security products for customers worldwide.
Download the paper in PDF format here.