Integration of Genetic Algorithm with Artificial Neural Network for Stock Price Forecasting
Anggia Suci Lestari

Department of Physics, Bandung Institute of Technology
Jalan Ganesha 10, Bandung 40132, Indonesia
anggiasucill[at]gmail.com

Abstract

Econophysics is a discipline that applies ideas, methods, and models in statistical physics and complexity to analyze data from economic phenomena. One of the objects to be addressed is the stock market. Approaches that can be used to model the economic sector are data analysis and physical models with computational physics. In this research, a predictive model for the closing price is created using the integration method between Genetic Algorithm (GA) and Artificial Neural Network (ANN), in this case Backpropagation (BP). The integration is then called GA-BP. GA is used to optimize the architecture and network weight values on BP structure so that prediction results will be more accurate. This research also analyzes the parameters and performance resulting from the model created. The data used in this research are the daily stock prices of AAPL (Apple Inc.), SPLK (Splunk Inc.), and BA (Boeing Co.) from December 31st, 2019 to December 31st, 2022. From this research, the integration model succeeded in producing a prediction model with better performance evaluation than using the BP model alone based on its MAE, MAPE, and R2. The integration model can also provide good accuracy in predicting stock movement patterns for the next few days.

Keywords: Artificial Neural Network, Backpropagation, Genetic Algorithm, Stock Prediction

Topic: Modelling and Computational Physics