RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2
Project

NevaMedia Gmbh (Deductor)

Customers: NevaMedia Gmbh

Moscow; Information technologies

Product: Deductor

Project date: 2014/01  - 2004/12

Content

Client's profile

NevaMedia Gmbh is the company which is engaged in professional translations and computational linguistics. NevaMedia provides qualitative translation, imposition, localization, prepress processing in all European, Russian and Turkish languages.

Task

In customer bases of data a large amount of materials with the description of parts and constructions collected. These descriptions were prepared by engineers during quite progressive tense and were formed them in any form. Such text is easily processed by the person, but badly gives in to computerized analysis. And the impossibility of application of the automated methods considerably increases the cost of its processing.

Creation of the mechanism allowing to clean and standardize texts with the description of parts and constructions was a project objective. Standardization is meant as process of extraction of 'article' (lines with the description something) significant data and their systematization according to the predeterminated template.

Solution

The developed system solves an assigned task, based on mechanisms of self-training and an associative memory. The specialist analyzes article, and a system automatically studies at these data to make standardization and "prompts" to the user the most probable option of breakdown. In process of accumulation of "experience" the quality of standardization improves and processing speed increases.

Result

Result of machining – the standardized text which is very convenient for using for preparation of booklets, directories, technical guidances and implementation of transfer.