| A large amount of data information is produced every day in our daily life,some of them can be useful and some of them may not.In order to process these data and extract the useful data,feature selection is a most particular method.Feature selection plays an important role in machine learning.Its main goal is to maximize the classification performance and minimize the number of features.The purpose is to remove irrelevant or redundant features from a given set of features for finding the most important information and reduce the data dimension.Feature selection is a NP problem,and finding optimal subsets is a key problem in feature selection.In order to solve the problem of feature selection,the main research contents are as follows:(1)To solve the problem of population diversity,a binary bat particle swarm optimization algorithm was proposed.The algorithm introduced binary particle swarm optimization algorithm into the local search of binary bat algorithm,and compared the original bats with randomly generated bats to optimize the population and improve the search performance.(2)A hybrid binary bat particle swarm optimization(PSO)algorithm is proposed,this algorithm combine binary bat algorithm(BBA)with the characteristics of the bats explore in space by echolocation,and treat each particle in the binary particle swarm optimization algorithm as a bat in the bat algorithm to make the algorithm converge to the global optimal solution in the searching area.The simulation experiment analysis and the experimental results show that the two proposed algorithms have better searching ability than the traditional optimization algorithm.This thesis proposed and studied the performance of the two algorithms.The UCI dataset is used to evaluate the performance of the two algorithms.The results obtained by the KNN classifier prove that the two algorithms proposed in this thesis contain the ability to search the optimal features combination in feature selection. |
PDF Full Text Request