An Improved Genetic Algorithm and Its Application in Classification
In this paper, based on a simple genetic algorithm and combine the base ideology of orthogonal design method then applied it to the population initialization, using the intergenerational elite mechanism, as well as the introduction of adaptive local search operator to prevent trapped into the local minimum and improve the convergence speed to form a new genetic algorithm. Through the series of numerical experiments, the new algorithm has been proved to be efficiency. we also use this new algorithm in data classification, select 5 benchmark datasets and the experiment results shown the new algorithm can get higher accuracy than k-nearest neighbor method.
Keywords: Genetic Algorithm, Optimization, Classification, K-Nearest Neighbor, Population.
Download Full-Text
ABOUT THE AUTHOR
Xuesong Yan
School of Computer Science, China University of Geosciences Wuhan, Hubei 430074, China
Xuesong Yan
School of Computer Science, China University of Geosciences Wuhan, Hubei 430074, China