Return Forecasting, Risk Modelling & Portfolio Construction in the Banking Industry of South Africa

Abstract

The integration of return forecasting, risk modelling, and portfolio construction is crucial for developing robust investment strategies. This research investigates the benefits of using advanced machine learning techniques, specifically Long Short-Term Memory (LSTM) and Support Vector Machine models, in predicting stock prices. Financial markets are dynamic and complex in nature and applying more advanced techniques yields more benefits on the predictive accuracy of models when using historical data. An essential part of these research is building multiple models with various statistical techniques and machine learning techniques to compare their performance. The best performing model based on selected performance metrics was chosen and used to predict future stock prices for portfolio construction. The findings of this research offer insights into the practical applications of advanced machine learning techniques for financial forecasting. By using advanced machine learning in return forecasting, we can assess associated risks comprehensively, and construct well-diversified portfolios which helps investors achieve superior performance and resilience in various market conditions. As financial markets evolve, ongoing research and development in these areas remains essential for achieving optimal investment outcomes and managing financial risks effectively