| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 496784 | Applied Soft Computing | 2009 | 9 Pages |
The fuzzy-driven genetic algorithm for sequence segmentation consists of a genetic algorithm whose objective function is driven by a fuzzy fitness finder. The genetic algorithm starts with an initial population of alternate solutions where each solution is a different partitioning of the sequence into segments. The algorithm uses adaptations of the standard genetic operators to reallocate the partitions so as to achieve optimal segmentations. A fuzzy fitness finding mechanism evaluates the fitness values of the evolving segmentations, taking into consideration the combined effect of multiple heterogeneous features that have been identified as governing factors for the formation of the segments. The relationships between segment elements can also be modeled by this novel approach of applying soft computing paradigms to the segmentation of multi-dimensional sequences. The algorithm developed in this work has been successfully implemented for gene sequence segmentation to predict groups of functionally related genes that lie adjacent on the genome sequences of bacterial genomes.
