Research on an integrated index prediction model based on RF-XGBOOST-ANN
Chunlei Ji · 2023
At present, the performance of each machine learning algorithm is different in the exponential prediction research, and they work independently without fusion, and cannot complement each other, resulting in low accuracy, generalization and stability of each machine learning algorithm. This paper aims to find a better prediction algorithm by comparison. In this paper, by constructing a prediction method based on Random Forest-XGBOOST-Artificial Neural Network and integrated learning based on BAGGING, MSE, R2and MAPE are used as test indicators. The conclusions are as follows: The RF-XGBOOST-ANN integrated algorithm has great advantages over single machine learning, with small error and accurate prediction. It stands out among many integrated prediction models and has broad application prospects. In this paper, the ensemble algorithm is used to improve the prediction accuracy. The prediction results are more accurate. The machine learning algorithms are effectively integrated, and the prediction performance is significantly improved.