Data Extraction Platform

DEP - is a universal commercial software solution that replaced manual data collection processes and existing systems that are currently used to collect data from various financial instruments and securities. 

Based on various research and development, DEP will automatically find the required numbers or text in a semi-structured electronic documents in real time. Thus, there is a link to the original document to ensure full transparency of data collection. With DEP, detailed information can be selected from various documents, such as:

• • Securities
• • Tax and Accounting Statements 
• • Other reports of enterprises and organizations

Data Extraction Platform can perform processing of documents submitted in various formats, such as:

• • ASCII text files 
• • E-mail message 
• • XML-documents
• • HTML-documents 
• • PDF-documents

Selected data are checked for quality and correctness, and then added to the existing database scheme. The next step, the data provided by the user in the format of XML (or in any other format defined technical specifications).

With links to original documents DEP to perform a simple conversion of each number or text value to its precise location in the original document - thus satisfying the demand of transparency of data collection.

DEP consists of several functional blocks that provide effective and coordinated system performance, such as Extraction Template Builder, Document Repository, Text Model Builder, System Administration, and other important components.

Document Repository

Document Repository module allows users to make documents in the system and prepare for the data collection process.

Text Model Builder

Annex Text Model Builder is designed to build mathematical models of data collection. During processing, document selection model performs the desired value for a particular business metric. This application allows the user to edit the decision tree, as well as to optimize some parts of the tree.

Data extraction from the text 

Models building for the data extraction