Abstract: Aiming at the problem of low-quality initial solutions and insufficient diversity, we proposed a random forest model that introduced the Logistic chaos mapping to improve the optimization of the sea horse optimization algorithm. Firstly, after improving the sea horse optimization algorithm, it was combined with the random forest algorithm to improve the discriminative accuracy of the classic random forest algorithm. Secondly, in order to  verify the  performance of new model, comparative experiment  was conducted by using  five models for four evaluation metrics. The experimental results show that the model has accuracy rate of  96.15%,  precision of 100%, recall rate of 92.31%, and  F₁-Score of 96.00%, which improves the performance of the  random forest method.

Key words: optimization algorithm, sea horse optimization algorithm, random forest, classification algorithm, parameter optimization