Academic Journal of Computing & Information Science, 2024, 7(8); doi: 10.25236/AJCIS.2024.070803.
Jiakang Ma1, Weipeng Xu2
1School of Traffic & Transportation Engineering, Central South University, Changsha, 410083, China
2School of Economics and Management, Tiangong University, Tianjin, 300387, China
Accurate time series forecasting is crucial for decision-making in fields like financial analysis, weather prediction, energy consumption, and intelligent transportation. Traditional models often struggle with non-linear and complex data. This study aims to improve forecasting accuracy by integrating Genetic Algorithm (GA) with Back Propagation (BP) neural networks. BP neural networks are powerful but face issues like local minima entrapment and network structure selection difficulties. GA is employed to optimize the weights and thresholds of BP networks, enhancing their generalization and robustness. Testing on a non-linear, non-periodic time series dataset, the GA-BP model showed significant improvements in prediction accuracy and stability. Performance metrics such as RMSE, R2, MAE, and MBE confirm the model's effectiveness. This hybrid approach not only boosts forecasting accuracy but also introduces innovative methods for addressing complex time series challenges, making a valuable contribution to the field.
Back Propagation Neural Network, Genetic Algorithm, Prediction Model
Jiakang Ma, Weipeng Xu. GA-BP-based Nonlinear Time Series Forecasting: Method and Applications. Academic Journal of Computing & Information Science (2024), Vol. 7, Issue 8: 15-20. https://doi.org/10.25236/AJCIS.2024.070803.
[1] Bawdekar A A, Prusty B R, Bingi K. Sensitivity analysis of stationarity tests’ outcome to time Series facets and test parameters[J]. Mathematical Problems in Engineering, 2022, 2022.
[2] Hart A G, Hook J L, Dawes J H P. Echo state networks trained by Tikhonov least squares are L2 (μ) approximators of ergodic dynamical systems[J]. Physica D: Nonlinear Phenomena, 2021, 421: 132882.
[3] Duan Y, Liu M, Dong M. A metric-learning-based nonlinear modeling algorithm and its application in key-performance-indicator prediction[J]. IEEE Transactions on Industrial Electronics, 2019, 67(8): 7073-7082.
[4] Yang J, Hu Y, Zhang K, et al. An improved evolution algorithm using population competition genetic algorithm and self-correction BP neural network based on fitness landscape[J]. Soft Computing, 2021, 25: 1751-1776.
[5] Lu Y, Li T, Deng Y. Predication of Water Pollution Peak Concentrations by Hybrid BP Artificial Neural Network Coupled with Genetic Algorithm[J]. Applied Artificial Intelligence, 2024, 38(1): 2341356.