NN-Tool creates data models based on neural networks. It is especially
suited for applications in process industries. NN-Tool is used for modelling
by customers like Bayer, BASF, Henkel and Merck.
- Modelling Efficiency, using extremely fast algorithms, optimal data
preprocessing, automatic model optimization and an intuitive user
- Modelling Precision by automatic structure optimization, i.e.
NN-Tool determines independently the optimum network structure and the
relevant input parameters.
- Taylor-made application systems. Eventually you would like to
solve problems with your models.
Prof. Dr. Frank Bärmann
Tel.: ++49 211 7489973
NN-Tool consists of a simulator for creating neural network models
and of an application module for interpreting the created models
(prediction and optimization respectively).
After provision of data the models can be created
very quickly, because NN-Tool takes over independently the scaling
transformations as well as the identification of complete data sets, which
are necessary for neural nets (NN). The created NNs are of the standard type
(feedforward net with one hiddenlayer), which has been proved to be optimal
in all practically relevant task requirements . After creating a model the
user has the possibility to execute evaluation- and optimization calculations
on the basis of the created model. The results are recorded in an Access data
base. Therefore also the usual data management
methods are available to the user (especially SQL).
application consists of the three
- Data conditioning
- Net training
- Net evaluation for prediction and optimization
NN-Tool is used by the following
- Lucobit AG
- Ticona GmbH
Get a free trial version of NN-Tool 2018 (english/german
edition). Just click below: