Database Search

Information is the lifeblood of competitive companies. Accurate and prompt processing of incoming information is the cornerstone of a successful business.

eFLOW’s Database (DB) Search module offers high-performance approximate-search technologies for online and offline searches, in addition to data quality solutions, for databases with millions of entries.

The highest accuracy levels of data recognition are provided by this high-performance approximate-searching and matching module. eFLOW’s Database (DB) Search module is based on complex learning algorithms, and enables information processing of multi-field documents over large databases, whilst increasing efficiency and saving time.

When using the DB search module, the results from the OCR engine, which were only partly recognised, can be quickly and accurately validated against information that already exists in the database, without halting the process. Therefore business critical data is transformed into business-ready digital content.

Even requests with spelling errors, incomplete data or bad scan results can still be automatically validated against the reference database which avoids costly manual corrections.
In addition, the provision of a correction assistant with a list of approximate matches from the database can improve manual correction efficiency significantly.

Proven technology

Since its release at the end of 2004, TIS has provided eFLOW’s Database (DB) Search module for:

• Applications that need validations with large or incorrect databases (over 1 million entries), such as address databases.
• Applications where speed and data accuracy are vital.

Unique Features:

• Provides quality results in milliseconds
• Queries extremely large databases (tens of millions of entries)
• Returns reasonable answers to queries, even when the query is only approximately equal to the keyword (due to spelling errors, bad scanning, incomplete scan, incorrect database entry)
• Language-independent
• Alias dictionaries allow the transformation of terms and names: for example München to Munich, Big Apple to New York
• No change to IT infrastructure
• Bad scan & OCR results can be automatically validated against the reference database, avoiding expensive manual correction
• Improving manual correction (completion) by giving the operator a list of approximate matches from the database

---

Information request form

First name *		Last Name *
Company *		Job title *
Country *		Phone *
Email *

Fields Marked with an asterisk (*) are required