I am working on a new concept, Wish-IT®.SE, for creating permaculture within the framework of ideas, research, development and innovation. A concept integrating automatic customization and adaptation to user needs by applied artificial intelligence methods. It is based upon data mining, pattern recognition, classification, resolution methods, standard protocols and standardized free software modules.
A cooperative research project with WHO building a framework for automatic adverse drug reaction signalling. We performed data mining on a large data base of suspected adverse drug reactions using Bayesian neural networks and advanced search algorithms. We searched for complex patterns between large sets of variables in large data bases with sparse information. These methods can find important patterns also when they are rare and are able to learn even from single events. The outcome was generation of early warning signals on previously unknown adverse drug reactions as well as finding new adverse drug reaction syndromes and combined drug effects. Resulting in the methods now used for adverse drug signalling, where the first are running sharp since 1998.
Even Earlier work:
Process modelling with the help of RBF (Radial Basis Functions) and Bayesian probabilistic networks. Temporal segmentation using Bayesian neural networks.
Boycot hardware and software products with proprietary undocumented formats for information exchange! These formats are unsafe as they make you insecure and dependent.
There are at least four categories of computer users: (observe, I wrote these lines several years ago (1996) when professional users who wanted to buy support did not have open source as an option, now that has completely changed since some years, and for the gamers open source is getting bigger)