The Praedea DEP automatically and efficiently extracts specific
data from multiple sources to feed proprietary research, trading
and credit risk (e.g., Basle II) models in real-time. Some
of the many advantages that the Praedea DEP provides financial
institutions, including investment banks, commercial banks,
fund management companies (including hedge funds), financial
advisory firms, credit ratings agencies, research houses and
stock exchanges, include:
Cost Savings: No longer pay for data you don’t
need. The Praedea DEP allows you to collect only the specific
data that you desire customized to your precise specifications.
Additionally, the Praedea DEP allows you to achieve more with
limited internal technology and R&D resources.
Data Extraction: Upon the processing of documents with
the Praedea DEP, data can be utilized in end-user applications
almost instantaneously. No longer wait hours, days or weeks
after filings are made to receive the data you desire.
Data Extraction Platform: The Praedea DEP extracts
predetermined numeric figures and text from any semi-structured
electronic document – whether US or foreign, in any
With the Praedea DEP, text-mining models in large part build
themselves after initial manual tagging, enabling the building
of models to extract thousands of terms quickly and efficiently.
Efficiency: The Praedea DEP and its text-mining models
can be easily customized into existing work flow and processes
on a global basis.
Types: Documents are inserted and organized into the
Praedea DEP for data extraction from virtually any source
including SEC filings, FTP sites and internal document management
and Accurate Data: Automated data collection utilizing
the Preadea DEP eliminates the inconsistencies and inaccuracies
associated with much of the data currently provided in the
marketplace. The Praedea DEP’s text-mining models are
continually optimized for best performance. To further enhance
accuracy, the Praedea DEP’s text-mining models also
apply “company-specific memory,” the ability to
recall how specific companies report information.
Unlike data you may receive from other sources, the Praedea
DEP provides you with the ability to create/utilize “enriched
data” (i.e., with audit-ability, commentary and custom
calculations) to enhance your end-user applications.
Control: Praedea’s data extraction process provides
for efficient quality control. Data validation occurs based
on your custom quality control rules and error-checking protocols
are run with minimal human interaction; manual overwrites
of the auto-extracted data are possible, if desired. As such
manual overwrites occur, the text-mining models are capable
of employing a self-learning refinement process to incorporate
such alterations in future data extraction processes.
Transparency: With data extracted by the Praedea DEP,
highly-paid professionals are no longer required to spend
valuable time manually verifying the accuracy of data. Source
Links, embedded upon extraction, provide hyperlinks from any
output format (including Microsoft®
Excel) to precise locations in underlying source documents,
facilitating efficient quality control and value-added end-user
Built on open standards, the Praedea DEP is broadly compatible
with most server and database platforms and other software
applications. Extracted data is deposited into existing database
schemas or delivered as XML or XBRL files for easy access
and utilization. End-user applications such as Microsoft®
Excel, Microsoft® Word,
Lotus Notes® or other custom
analytic applications may be populated with the extracted
Implementation: Basic installation of and on-site training
regarding the Praedea DEP are typically completed in only
one day with minimal engineering resources. No prior programming
or text-mining experience is required. In most cases, initial
text-mining term models are functional within days, with full
automation-assisted production occurring within a month.