On-Line Feature and Classifier Selection for Agricultural Produce

S. Laykin, Y. Edan, and V. Alchanatis (Israel)


Unsupervised classification, classifier selection, fuzzy rule-based system and feature selection


This paper presents an on-line hierarchical classifier for agricultural products. The classifier consists of two levels. The first level detects new populations using an on-line clustering algorithm. The second level selects the best-fit classifier using a fuzzy system. This paper presents the combination of the two levels into a complete system. Feature selection is conducted on-line according to the classified population. A synthetic dataset is used to estimate the classifier capabilities and compare it to previous results. Results indicated that the combined on line system results in improved classification accuracy.

Important Links:

Go Back