This program was developed as tool for solving of two tasks:

• the recognition of associative links between fields of records of Database, and learning of Neural Network for use of these links for find of most probably values of fields in determined context,
• the forecasting of fields of next record of Database if the records of Database are points of any process.

The source of Database for Neural Network is SQL-query, supported by Borland Data Engine. This version of program (v. 3.0) uses Hopfield model of Neural Network, created as DLL. The project of trained Neural Network and DLL may be used in any applied program for forecasting or associative search in any area.

This program is developed in Delphi 3.0.

Now the russian next version (v. 3.1) is available. New opportunities are in this version:

• the appending to SQL-query of the field for detecting of trend and use one for forecasting,
• the masking of determined fields in SQL-query for Neural Network,
• print and save report of forecasting.

New version (v. 4.0) is developing now. This version will use the some models of Neural Networks for solving tasks: associative research of fields, forecasting of fields and clustering of records of Database.