B. Kusumoputro and Z. Rustam (Indonesia)
Fuzzy-LVQ, eigenspace, artificial odor recognition system, quartz micro balance
There are still major difficulties in the usage of fuzzy neural networks based on LVQ (FNLVQ) algorithms, i.e choosing the initialization of the fuzzy-reference vectors. The initialization step is important due to different selections of the initial reference vectors may potentially lead to different partition for different classes, which hampered the superiority of the algorithm. In this paper, we proposed a novel initialization method, by transforming all data from the origin problem space into its eigen space prior the usage of FNLVQ. Experiments are conducted in an artificial odor recognition system and it shows that the performance of FNLVQ in eigen space has higher recognition rate compare with that of in aroma space, especially for 18 class of three-mixture odors problem.
Important Links:
Go Back
