Automated/Manual Parameters Selection
gives users the flexibility to automatically or manually adjust the
network parameters for analyzing data. In auto mode, the efficient
learning algorithm will automatically adjust network parameters to
find a suitable architecture. In manual mode, the user is able to
manually fine-tune the parameters.
Rule Extraction and Pruning Strategy
provides rule explanation capability to gain a wider degree of user
acceptance on the results and to let user understand the potential
and ability of the trained engine in handling classification
problems. Pruning strategy enables user to remove low confidence and
unimportant prototypes (knowledge) from the system to reduce
complexity and network size.
Retrieve Old Knowledge
enables user to retrieve old prototypes from the database and
combine them with new input samples to form new prototypes. Thus,
the time used to retrain is significantly reduced.
Multiple Classifier Systems enable MetierView Predictive Analysis to
produce better results and higher accuracy in decision making using
a voting system mechanism.
can accommodate different category of data sets (e.g. data on credit
card spending, health diagnosis, manufacturing operation and etc) in