An Improved Artificial Bee Colony Algorithm and Its Application in Machine Learning

Abstract

In order to effectively overcome the problems and shortcomings of the basic improved artificial bee colony algorithm in solving the optimization accuracy of complex standard test functions, the low accuracy of the solution function and the large blindness of the local search strategy, a more basic improved artificial bee colony is proposed. Parameter optimization algorithm. The above artificial bee colony algorithm introduces the conjugate gradient method with strong local optimization search strategy performance in the follow bee stage of the basic improved artificial bee colony algorithm to change the optimization search strategy, and replaces the blind search strategy with a local deterministic optimization search strategy. The sexual optimization search reduces the randomness, enhances the local certainty of the food source of the follower bee and the optimization search ability, and ensures that each update of the follower bee’s food source will quickly be substantially improved. This improved artificial bee colony algorithm is widely used in parameter optimization of traditional dengue virus propagation model. The simulation results of the improved standard test function on the problem show that the improved artificial bee colony parameter optimization algorithm has higher optimization and solution function accuracy during simulation than the basic improved artificial bee colony parameter optimization algorithm. The obtained standard test the parameter output corresponding to the dengue virus model parameter output is in good agreement with the actual situation when the data is simulated.