We create specifically tailored versions of our models for you - our tools are self-adaptive and can be improved easily with only a limited number of in-house measurements. As an example : on a test set of 95 809 Pfizer molecules, it has been shown that by application of the self-learning feature of our software the mean prediction uncertainty can be reduced 1.02 log units to 0.59 log units (Tetko et al, 2009) .
You do not have to provide us your structure data for this service. We have experience handling encrypted data in the model creation process and you can execute the steps for which a knowledge of the structural information is necessary with the help of our software easily and safely at your place.
By using our database and development platform together with our our self-learning technology, the adaptation of existing models can be carried out quickly and inexpensively.
We create your models and conduct literature, database searches and data migration for you in no time. Get an overview of the information content of your data and a clear view of data quality, inconsistencies, and the achievable prediction accuracy.