Application of Feedforward Neural Network in Portfolio Optimization and Geometric Brownian Motion in Stock Price Prediction

This research study intends to predict the stock price movement of a set of stocks using the Geometric Brownian Motion and optimize the portfolio built from selected stocks using Feed Forward [1] Neural Network. The onset of the current global pandemic has led to major negative implications for the overall economy and personal financial situations. With many people losing jobs, getting reduced salaries, and exhausting their savings, they have started to invest heavily. The Securities and Exchange Board of India (SEBI) reports that there were 4.9 million new Demat accounts opened in the fiscal year 2020, which was the biggest in at least ten years. Thus, the study of the Stock Exchange is very relevant to the current times. We shall explain the importance of learning about investing in the Stock Market and derive an efficient model to make real-time predictions of a stock price and build an optimized portfolio. In the finance sector, portfolio optimization is a critical responsibility. The goal of portfolio optimization is to increase return on investment while lowering portfolio risk. Due to its capacity to understand intricate correlations between input and output variables, feed-forward neural networks have been successfully used to solve portfolio optimization issues. We explore the research on portfolio optimization using feed-forward neural networks in this paper.