Real-Valued GCS Classifier System

Abstract

Learning Classifier Systems (LCSs) have gained increasing interest in the genetic and evolutionary computation literature. Many real-world problems are not conveniently expressed using the ternary representation typically used by LCSs and for such problems an interval-based representation is preferable. A new model of LCSs is introduced to classify realvalued data. The approach applies the continous-valued context-free grammar-based system GCS. In order to handle data effectively, the terminal rules were replaced by the so-called environment probing rules. The rGCS model was tested on the checkerboard problem.

Modified Himmelblau Function Classification with rGCS System

“…While working with simple training sets (like the checkerboard problem investigated in [3]) it is optimal to set the quantity of the EPRs equal to the number of classes the input space is divided into. In more sophisticated problems as long as the input space is single dimensional or the number of divisions in each dimensions stays the same we can still use one population of environment probing rules.…”

“…In [3] we introduced a new model of real-valued LCS -the rGCS -to classify real valued data. rGCS is based on Grammar-based Classifier System (GCS), which was used to process context free grammar (CFG) sentences [5].…”

Towards 3D Modeling of Interacting TM Helix Pairs Based on Classification of Helix Pair Sequence

“…Both EM and GP approaches managed to, respectively, learn probabilities of Stochastic Context-Free Grammars (SCFG) for RNA structure prediction [15,16,17] and derive non-probabilistic CFGs for non-biological problems [18]. Successful applications of evolutionary algorithms to SCFG [19,20,21] include our earlier research on SCFGs for protein binding sites [22]. Since, unlike EM techniques, GA-based grammar inference allows introducing pressure towards more compact grammars (see Methods) and is less dependent on initial estimates of rule parameters [20], we choose this approach for learning grammar rules.…”