Artificial Neural Network-Based Forecasting to Anticipate the Indian Stock Market

The introduction: One of the most challenging situations in today’s stock market is predicting stock values. The stock price is a forecast time series that is becoming an enormously problematic task due to its features and dynamic nature. Case description: Stock prices and movements are frequently predicted using artificial neural networks (ANN) and Support Vector Machines (SVMs). Every algorithm has a different method for predicting and learning patterns. The ANN is a standard tool for producing global financial predictions that incorporate research and analysis. Discussion and evaluation: Backpropagation Neural Network (BPNN), Support Vector Regression (SVR), and Support Vector Machine (SVM) are by far the most prominent approaches used in financial time-series prediction. In this paper, we examine the outcomes of neural networks consisting of five distinct learning algorithms, namely Bayesian Regularization (BR), Levenberg–Marquardt (LM), stochastic gradient descent (SGD), Adaptive Momentum (Adam), and Scaled Conjugate Gradient (SCG), for stock price forecast using valid dataset Indian stock market. Conclusion: All five algorithms achieve an accuracy rate of 98.5% when applied to a dataset of the Indian stock market. The accuracy for Bayesian Regularization (BR), SCG, SGD, Adam, and LM 97.01%, 95.0%, 96.12%, 97.99%, and 95.3%, respectively, is much lower than the findings achieved using valid data.