Assessing ML models performance predicting Maltese stock movements during local government elections

Abstract

Predicting the stock market's response to governmental elections has historically been a challenge in financial research, due to the complex connection between political events and investor behaviour. This dissertation looks at the potential influence of electoral results on the Maltese stock market, specifically aiming to find patterns for market fluctuations during these pivotal political events. The work integrates data analysis and machine learning approaches to assess the efficacy of several predictive models in capturing the impact of elections on stock trends, providing a novel insight into politically influenced market behaviour. A specialised dataset was developed to focus on stock performance during election seasons. Each data point is defined by unique attributes, including fluctuations in trading volume, price volatility and sector specific influences. These aspects are intended to illustrate the diverse responses of the market to electoral results. Each case is thereafter categorised according to whether the Malta Stock Exchange Index increased or decreased following the electoral event.