Technology Overview

The challenge facing researchers today is how to harness the potential knowledge in billions of documents—in scarce time. Most research tools available today rely on domain-specific “taxonomies” that cover a limited range of data sources and research questions. And all rely on keyword search, which too often delivers millions of documents that users need to read.

Using state-of-the-art text analytics technology, NetBase has developed a proprietary, patent-pending technology called the NetBase Research Engine that powers all of our solutions. NetBase Research Engine was built to go beyond keywords to find and extract comprehensive and precise answers, reducing the time spent on irrelevant results.

The Problem with Keywords

Keyword-based search engines are primarily designed to return documents matched by a set of keywords, ranked by word frequency, proximity and popularity. This approach is far from ideal for many types of research tasks because:

(1) What is sought is rarely popular or well known.
(2) Many ideas are often expressed in ways not easily found using keywords.
(3) Getting a complete answer requires hours, days, or even weeks of browsing, reading, and analyzing countless documents.

NetBase Research Engine

NetBase Research Engine uses state-of-the-art text analytics technology to search, extract and summarize information with rich interrelationships from billions of business-relevant web sites and scientific documents. The Research Engine is based on the structure, content, and meaning of sentences, not just keywords, and is designed to extract and link semantically related concepts, ideas, and entities.

NetBase's Semantic Index

The research engine runs off a semantic index, the world’s largest natural language database that contains billions of interrelated concepts, ideas, and entities extracted from vast amounts of scientific, technical, and business related information, including:

Billions of business-relevant web pages
Tens of millions of full-text scientific articles and abstracts
Tens of millions of patents from the top world-wide patent offices

The semantic index database contains entities like:

Products (products and services)
Business entities (corporations, institutions, organizations, experts, etc.)
Technologies (devices, tools, systems, material, chemical compounds, etc.)
Approaches (methods, processes, procedures, techniques, treatments, means, etc. )
Solution characteristics (capabilities, benefits, problems addressed, applications, uses, beneficiaries, pros, cons, etc.)
Market characteristics (market needs, problems, causes, victims, etc.)

The picture below illustrates the rich set of semantically-related entities and concepts extracted from a single sentence.