COVID-19's influence on Karachi stock exchange: A comparative machine learning algorithms study for forecasting.

Area of Science:

• Economics
• Data Science
• Financial Markets

Background:

• The COVID-19 pandemic significantly impacted global economies and financial markets.
• Understanding stock market behavior during health crises is crucial for economic stability.

Purpose of the Study:

• To identify the most effective machine learning (ML) model for predicting the Karachi Stock Exchange (KSE) performance during the COVID-19 pandemic.
• To analyze the interconnection between COVID-19 metrics and stock market fluctuations in Pakistan.

Main Methods:

• The study period spanned from March 1, 2020, to November 26, 2021, covering the peak COVID-19 period.
• Five machine learning models (Linear Regression, K-Nearest Neighbors, Random Forest, Regression Tree, Support Vector Machine) were applied to KSE 100 index data and COVID-19 variables.
• Model performance was evaluated using Mean Absolute Percentage Error (MAPE), Mean Squared Error (MSE), Mean Absolute Error (MAE), and R-squared (R²).

Main Results:

• The Random Forest (RF) model demonstrated superior predictive accuracy, achieving an R-squared value of 0.91.
• This finding contrasts with previous research suggesting a negative impact of COVID-19 on major stock markets.
• The study highlights the potential of ML models in navigating market volatility during pandemics.

Conclusions:

• The Random Forest model is recommended for predicting stock market movements influenced by pandemic-related factors.
• Insights can guide investors in strategic decision-making and policymakers in mitigating economic impacts.
• Further research may explore other financial markets and machine learning techniques.