Document mining system

Document mining system is represented by the demo-stand AMS for DED (Automated Mivar System for Detecting Errors in Documents).


Performance of any company can be represented as a combination of business processes. At each stage of business-process divisions of the company document information, collect statistics, compile reports on the results of work. These reports may contain errors of different types caused by human factor: errors in design, configuration, calculations, data conversion, false information, fraud. These errors may result in leakage of considerable amount of company’s legal revenue.

The system for detecting errors in documents was designed to prevent leakage of telecommunications company’s revenues. According to the requirements of the company, it was necessary to develop a system for detecting errors in the accounting documents concerning the lease of telecommunications equipment and estimating the cost of these errors. The logic, which the system is based on, can be applied to any cases related to control of revenue chain. The introduction of the system will ensure automation of monitoring and incident response processes.


The system should be capable of analysing statistical documents, detecting errors in them and calculating losses caused by these errors. The system should also have a convenient user interface for downloading documents and representing results of the analysis.


  • Downloading statistical documents and checking for template validity.
  • Analysis of downloaded documents for errors.
  • Calculation of cost of all the errors and showing the result of the analysis to the user.