House Price Prediction Model Based on Multi-Strategy Improved COOT Optimization Algorithm

The COOT optimization algorithm holds promise for optimizing machine learning algorithms and feature selection. However, its drawbacks, such as weak overall search capability and sometimes it goes to a local optimum, pose challenges when applied in practice. Aiming at the problems existing in the reference white crown chicken algorithm, a new algorithm named TLCOOT is proposed, which is based on tent Chaos Mapping and Levy Flight Strategy. This strategy changes and improves the strength of the original algorithm to jump out of partial optimal solutions. A variety of functions are used to evaluate the effect of the algorithm. The results show that TLCOOT is superior to the COOT optimization algorithm and several swarm intelligence algorithms in Best, Mean, Std and convergence curve. On this basis, TLCOOT optimized random forest house price prediction model (TLCOOT-RF) was used. The experimental results of the model verify the practicability of the proposed TLCOOT optimization algorithm.