Web based predictive defuzzifier
Defuzzification is a method of converting membership functions generated from linguistic variables into usable numerical values. Defuzzification is an important process in the fuzzy inference engine but in its present state, there are more than 50 (fifty) deffuzication method however, there are no methods to help known method for linking a defuzzification method with a particular type of data. Our system will allow users to upload their data in .xls and .csv and get it processed by the defuzzifier, who would have predicted the best defuzzification method for the class of data recived. The systems build knowledge from the submitted data by learning and using it for training in other to perform better while processing similar data.The work will leverage on the the open science using FG, Kepler, OLAP cube to build the defuzzifier and use PHP for the front end to enable users upload data for processing.
The use case is still in an early stage of its implementation. Plans are, however, to develop a portlet to be included in the Africa Grid Science Gateway and submit jobs to the undelying e-Infrastructure using the instance of the FutureGateway API Server – developed in the context of the INDIGO-DataCloud project – which is connected to the Science Gateway.