Headquartered in Silicon Valley, Glenbrook Networks was founded in 2001 with the objective to deliver a next generation search technology, one that would enable the extraction of information from the wide-open sea of online information with a high degree of precision. The company has developed over the past seven years a unique technology platform that trawls surface and deep web, analyzes web pages, and automatically extracts unstructured data.

With the constant increase in web page complexity and heterogeneity, and with steadily rising average number of keywords in queries (2.3 in 2005 vs. 4.2 in 2007), it has become increasingly difficult to find relevant results using traditional search mechanisms. Web pages contain multiple "contexts" (articles, posts, ads, tables of contents, links to other pages, etc). The more keywords in a user query, the more often they end up being dispersed among these "contexts" instead of actually appearing together in one single context. This makes an apparent match completely irrelevant. Traditional proximity techniques fail to correct for this. Furthermore, even if the page is truly relevant, it takes time for a user to find the portion that is relevant to the query.

The core of Glenbrook's platform is a page analysis system that evaluates web pages and identifies prominent contexts. The human brain can easily identify the individual components of a web page and evaluate them for importance and relevancy. Yet, a machine-based page segmentation and analysis system that mirrors this simple human activity is difficult to build. Automation of this approach has been a technically difficult problem to solve, as evidenced by the absence of such a solution in today's popular search engines. The complexity of the problem is comparable to, but more difficult than, the challenges previously confronted by experts in optical character recognition (OCR). HTML and, especially, DHTML pages are effectively multi-dimensional structures that include dynamically rendered texts, images, and inclusions. The practically unlimited flexibility of DHTML language significantly increases the complexity of algorithms needed to recognize page contexts.

The founders of the Company have committed years of work, research and experience to the development of the algorithms, architecture and technology that underpin the capabilities of Glenbrook Networks. Their unique set of combined skills, spanning Pattern Recognition, NLP, Search and Information Retrieval, Computer Chess, Databases, and Graph Theory, has allowed them to design a hybrid system that combines a number of different technologies, in such a way as to eliminate the limitations of each method.

Oh, and why are we using Glendor as the name of our service? Glendor is sort of a contraction of "Gleaning d'Or", implying that hopefully we will help our users glean gold nuggets thanks to our system's ability to analyze web pages and extract relevant "contexts".

Contacts