Ontos SOA

Ontos SOA realizes the whole cycle of semantic-syntactical processing, management and analysis of unstructured information located in the Internet and large corporative data banks.

The method of analyzing and handling unstructured information can be viewed as a process consisting of the following consecutive steps:

1. Monitoring of new media messages.
2. Linguistic analysis of message texts.
3. Determination and analytical processing of identical objects (persons, organizations, political parties. etc.), with subsequent saving of data in a dedicated repository (data warehouse).
4. Provision of document meta-information processing services.

Linguistic analysis of text

	Using natural language processing with linguistic rules, the system recognizes objects and relations in a text. The objects and relations of interest are specified in a domain ontology. The recognized and extracted information can be organized in a computer-readable form, such as tables, and stored in a data base. Thus, enabling users to work with it as with any other database.

It is important to mention an additional step of saving the obtained document meta-information in our semantic data store. The step of identifying and merging of duplicate objects.

For example, suppose we identified a fact stating that "Paul Wolfowitz is a World Bank president". When storing this information the system finds that there is already some information presented in a database for example "From 2001 to 2005 Paul Wolfowitz worked as Deputy Secretary of Defense". So, all we need to do is just add additional information about the person Paul Wolfowitz and extend the knowledge.

Distributed computing

Due to the resource-intensive nature of most operations, the process can be executed on several computers simultaneously.

The developed system is built with the latest j2ee technologies and allows the installation on large servers or on common desktop PCs.