Abstract: In order to improve the accuracy of feature selection of bat algorithm， a global chaotic bat optimization algorithm(GCBA) was proposed. Firstly， GCBA adopts chaotic mapping method to enable the initialization of the population to traverse the entire solution space， obtain the optimal position of the bat， and make it more abundant. It solved the problem of initial population randomness. At the same time， GCBA introduces the optimal solution of the current particle and the optimal solution of the current population to jump out of local optimal solutions， which can effectively avoid the premature and improve the global search ability of the algorithm. The results of the bat algorithm(BA)， particle swarm optimization(PSO) and genetic algorithm(GA) on 10 data sets showed that the proposed algorithm has higher classification accuracy and stronger ability to jump out of local optimum.

Key words: bat algorithm, chaotic mapping , global optimum, local optimum, feature selection