PROJECT TITLE:

Ensemble of Adaptive Rule-Based Granular Neural Network Classifiers for Multispectral Remote Sensing Images - 2015

ABSTRACT:

Data granulation opens ample scope to design probably clear neural networks known as granular neural networks (GNNs). The paper proposes a classification model in the framework of ensemble of GNN-primarily based classifiers, and justifies its improved performance in classifying land use/cowl classes of multispectral remote sensing (RS) pictures. The model additionally provides an adaptive method for fuzzy rules extraction from the fuzzified input variables for GNN and therefore avoid the uncertainty in empirical search of rules for output category labels. The superiority of the proposed model to other similar methods is established both visually and quantitatively for land use/cowl classification of multispectral RS pictures. Comparative analysis revealed that GNN with multiple rules performed higher than GNN with single rule assigned for every of the categories, and ensemble of GNNs outperformed all other strategies. Varied performance measures, like overall accuracy, producer's accuracy, user's accuracy, kappa coefficient, and live of dispersion estimation, are used for comparative analysis.