Prediction of House Price in Boston Based on the Fusion of Crested Porcupine Optimizer Algorithm and BP Neural Network

Aiming at the problems that the traditional backpropagation (BP) neural network is prone to fall into local optimum and has a slow convergence speed in house price prediction, this paper proposes a BP neural network fusion model (LCPO-BP) based on the improved crested porcupine optimizer algorithm. By introducing the elite reverse learning strategy and the adaptive weight mechanism, the global search ability is enhanced, and the initial weights and biases of the network are optimized, thereby improving the performance of the model. The experimental results based on the Boston house price dataset show that compared with the traditional BP, the LCPO-BP model reduces the mean absolute error (MAE) and root mean square error (RMSE) by 58.81% and 59.63% respectively, and the coefficient of determination (R²) increases to 0.93269. The effectiveness of the improved algorithm in avoiding local optimum, accelerating convergence and improving prediction accuracy was verified, providing a new reference for intelligent optimization methods for complex nonlinear prediction problems.