The Application of a Backpropagation Neural Network for the Prediction of the New York Stock Exchange

Stocks have significantly impacted the market over the past century. Nevertheless, the domain of stock market prediction has consistently been perceived as a prospective yet predominantly ineffectual discipline. At the same time, artificial intelligence networks offer distinct advantages in data processing. Numerous scholars are convinced that artificial intelligence networks are instrumental in enhancing the accuracy of stock market predictions. Furthermore, they have effectively implemented sophisticated algorithms, including Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs), in the domain of stock market forecasting. In this study, we employ a backpropagation neural network to predict stock prices on the New York Stock Exchange, evaluating its accuracy across different neural network node configurations and datasets. Upon inputting data varying in magnitude, we yielded mean squared errors (MSEs) corresponding to diverse scales. In summary, we successfully predicted stock market trends, identifying instances of overfitting and underfitting, as well as the conditions for precise forecasting.