CA For EBM

Computer-Aid For Experience-Based Modeling


> Like SBM became a world power due to science and computer development, EBM also can be enhanced, from 2 to 5 parameter models, that can be handled directly by the human brain, to networks of multi-dimensional real-world EBMs with 20 to 50 parameters or even more. The key is bringing Computer Aid to EBM: CA+EBM=CAEBM !


> CA can be used to collect, refine, and handle the often big number of examples from arbitrary sources, over appropriate normalizations and transformations in accordance to the list of OPs and IPs, up to correct outliers and filling data gaps of "missing values" etc.


> CA is indispensable to setup, train, and control the appropriate high-dimensional KHB, representing the complete relationships between the IPs and the OPs, in accordance to the problem at hand. Within this, generalization and QA methods have to be applied carefully, to ensure the accuracy and quality of the EBMs, and of it's results.


> CA finally is used to deploy the EBMs in real-world environments, and to give confidence measures to any single result in the high-dimensional problem space, based on the local availability of examples, which support the model at hand. 


> Along this line, CA can enhance the restricted EBM capabilities of the human brain to a general and powerful technology, to tackle also complex inter-disciplinary problems, which are not even covered by today's sophisticated and high-developed SBMs. This way, new opportunities to win and to apply the know-how contained in myriads of existing examples arise.