Multiple composition ceramics synthesis data was analyzed by Artificial Intelligence Algorithms were Inductive learning Algorithm(C4.5) and Back-propagation Neural Network Algorithm(Bpn5).
For rule extraction C4.5 Inductive learning Algorithm made by J. Ross Quinlan was used. And for the evaluation of predictability on the trained Neural Networks, Goodness of Fitness testing was used.
The Goodness of Fitness testing results of the predicted data by Bpn5 was compared with the predicted data by SAS multiple regression.
The C4.5 inductive learning algorithm extracted hidden rules from the ceramics synthesis data sets and the rules had 2 dimensional decision features.
The predictability of the trained Neural Network evaluated with Goodness of Fitness testing was much better than the SAS multiple regression.
