PMML, the Predictive Model Markup Language, allows for a predictive analytic model to be developed in one application and easily moved to another for production deployment and execution.
Once a predictive model is exported from a PMML-compliant tool such as SAS EM, SPSS/IBM, R, KNIME, RapidMiner, ... it can be uploaded directly into the Zementis ADAPA engine which makes the model available for execution via its console or as a web-service. ADAPA can already import most of the techniques defined by the PMML standard and now, with the release of ADAPA 3.4, we have expanded it even further to cover Association Rules.
Analysts always want to explore rules and relations between variables in large data sets. The learning mechanism of Association Rules serves this purpose. The rules discovered by Association Rules often provide useful information for marketing activities. For example, they can be used for discovering relations between products in transaction data in supermarkets. In this way, an association rule can be found to indicate that if a customer purchases beef and cat food together, he/she is most likely to also buy tuna cans.
An Association Rule Model in PMML is represented by the element "AssociationModel". ADAPA and PMML support two different formats for representing Association Rules. These are "rectangular" and "transactional". To learn more about these two formats, please read our posting: Association Rules in ADAPA.
In addition, with the release of ADAPA 3.4, we were able to make ADAPA even better when it comes to converting and correcting PMML files. This is yet another big step towards true interoperability. In many cases, even if the model has syntactic or semantic problems, ADAPA automatically corrects known issues for models exported from several model development environments. For that, we analyze PMML files submitted to us by our partners and clients.
If for any reason, your PMML code cannot be converted or corrected automatically, feel free to contact us. We are here to help!