Meta-cognitive RBF Network and its Projection Based Learning algorithm for classification problems

• Published in: Applied soft computing Vol. 13; no. 1; pp. 654 - 666
• Main Authors: Babu, G. Sateesh, Suresh, S.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.01.2013
• Subjects: Meta-cognitive learning; Multi-category classification; Projection Based Learning; Radial basis function network; Self-regulatory thresholds; Projection Based Learning; Meta-cognitive learning; Radial basis function network; Multi-category classification; Self-regulatory thresholds
• ISSN: 1568-4946; 1872-9681
• DOI: 10.1016/j.asoc.2012.08.047

[Display omitted] ► Meta-cognitive learning to emulate human learning components such as what-to-learn, when-to-learn and how-to-learn from sequence of training data. ► Sample learning, sample deletion and sample reserve strategy are proposed. Meta-cognitive component in PBL-McRBFN choose of the strategy based on information present in current sample and existing knowledge in RBF. ► PBL-McRBFN evolves the network architecture automatically and the strategies are also adapted to accommodate coarse knowledge first followed by fine tuning. ► Sequential learning algorithm uses computationally less intensive project based learning algorithm. ► Performance of proposed algorithm is compared with well-known fast learning neural networks reported in the literature using UCI data sets. ‘Meta-cognitive Radial Basis Function Network’ (McRBFN) and its ‘Projection Based Learning’ (PBL) algorithm for classification problems in sequential framework is proposed in this paper and is referred to as PBL-McRBFN. McRBFN is inspired by human meta-cognitive learning principles. McRBFN has two components, namely the cognitive component and the meta-cognitive component. The cognitive component is a single hidden layer radial basis function network with evolving architecture. In the cognitive component, the PBL algorithm computes the optimal output weights with least computational effort by finding analytical minima of the nonlinear energy function. The meta-cognitive component controls the learning process in the cognitive component by choosing the best learning strategy for the current sample and adapts the learning strategies by implementing self-regulation. In addition, sample overlapping conditions are considered for proper initialization of new hidden neurons, thus minimizes the misclassification. The interaction of cognitive component and meta-cognitive component address the what-to-learn, when-to-learn and how-to-learn human learning principles efficiently. The performance of the PBL-McRBFN is evaluated using a set of benchmark classification problems from UCI machine learning repository and two practical problems, viz., the acoustic emission signal classification and the mammogram for cancer classification. The statistical performance evaluation on these problems has proven the superior performance of PBL-McRBFN classifier over results reported in the literature.