In a world characterized by increasingly complex financial markets, the prediction of financial crises is a constant challenge. This bachelor thesis investigates the use of machine learning, in particular regression algorithms, to analyze and predict financial crises based on macroeconomic data. By building six different regression models and optimizing them using cross-validation and GridSearch, the feasibility of using these technologies for accurate predictions is discussed. Although traditional models show limited effectiveness, the integration of machine learning, especially kNN algorithms, reveals significant potential for improving prediction accuracy. The paper highlights the importance of classification algorithms and provides crucial insights for application in real-world scenarios to provide valuable tools for policy and business decision makers.
Inhaltsverzeichnis (Table of Contents)
- Introduction
- Related Work
- Empirical Evidence for Financial Crises
- Definition and Classification of Financial Crises
- Currency Crises
- Sudden Stops
- Debt Crises
- Banking Crises
- Financial Crises of the Past
- The Great Depression of 1929 to 1939
- First stage: A Flux in Foreign Exchange Markets
- Second Stage: Some Shifts in the Volume and Direction of International Lending
- Third Stage: A Rapid Institutional Change in the Banking System
- The Great Depression and the Friedman-Schwartz Hypothesis
- The Global Financial Crisis of 2007 to 2009
- First Factor: Expansive Monetary Policy
- Second Factor: Flawed Financial Innovations
- Third Factor: The Collapse of Trading
- The Great Depression of 1929 to 1939
- Definition and Classification of Financial Crises
- The Data Analytics Process
- (Traditional) Approaches to Predict Financial Crises
- Linear Models
- OLS Regression
- Ridge Regression
- Support Vector Machines (SVM)
- Tree-based Approaches
- Decision Trees
- Random Forest (RF)
- Neural Network (NN)
- k-Nearest Neighbors (kNN)
- Linear Models
- Statistics vs. Machine Learning
- Best Fit vs. Generalization: Risk of Overfitting or Underfitting
- Cross-Validation & GridSearch
- Measuring the Quality of Fit (MSE)
- (Traditional) Approaches to Predict Financial Crises
- Methodology
- Dataset Retrieval and Description
- Data Preparation
- Exploratory Data Analysis
- Detailed Feature Analysis
- Collinearity
- Handling Outliers
- Handling Missing Values
- Partitioning the Data
- Normalization
- Exploratory Data Analysis
- Constructing the Regression Algorithms
- GridSearch and Cross-Validation
- Report Data Frame
- Evaluating the Regression Algorithms: The Prediction Power
- OLS Regression
- OLS Regression's Prediction Power
- Ridge Regression
- Ridge Regression's Prediction Power
- Support Vector Regression
- SVM Regression's Prediction Power
- Random Forest
- OLS Regression
- The role of machine learning in predicting financial crises
- The effectiveness of various regression models in forecasting financial downturns
- The application of machine learning algorithms on macroeconomic data
- The potential of proactive decision-making in mitigating the impact of financial crises
- The challenges and limitations associated with predicting financial crises using machine learning
- Introduction: This chapter introduces the concept of financial crises and their impact on economic stability. It highlights the potential of machine learning methodologies for handling big data and predicting crises. The research objective and methodology are outlined.
- Related Work: This chapter explores previous research on financial crises and machine learning techniques. It provides a comprehensive overview of relevant studies and identifies gaps in existing knowledge.
- Empirical Evidence for Financial Crises: This chapter examines the historical occurrences of financial crises, providing a theoretical framework for understanding their causes and characteristics. It delves into the classifications of currency crises, sudden stops, debt crises, and banking crises. Specific examples, such as the Great Depression and the Global Financial Crisis, are analyzed in detail.
- The Data Analytics Process: This chapter discusses the methodologies employed in predicting financial crises, both traditional approaches and machine learning techniques. It explores the strengths and limitations of various models, including linear models, tree-based approaches, and statistical methods.
- Methodology: This chapter outlines the data collection and preparation process, including exploratory data analysis, outlier handling, and data partitioning. It describes the construction of the regression algorithms, highlighting the implementation of GridSearch and cross-validation.
- Evaluating the Regression Algorithms: The Prediction Power: This chapter presents the results of evaluating the constructed regression models, examining their prediction power and effectiveness in forecasting financial crises. It analyzes the performance of OLS regression, Ridge regression, Support Vector Regression, and Random Forest.
Zielsetzung und Themenschwerpunkte (Objectives and Key Themes)
This research explores the feasibility of predicting financial crises using machine learning algorithms on macroeconomic data. It aims to assess the effectiveness of various regression models in forecasting economic downturns and to evaluate their practical applicability in real-world scenarios. This research seeks to contribute to the understanding of financial crises and to provide policymakers and stakeholders with a tool for proactive decision-making.
Zusammenfassung der Kapitel (Chapter Summaries)
Schlüsselwörter (Keywords)
This research focuses on the application of machine learning algorithms, including regression models, in the prediction of financial crises. Macroeconomic data is used to train and evaluate the models, exploring their ability to forecast potential economic downturns. Key terms include financial crises, machine learning, regression algorithms, macroeconomic data, prediction, and forecasting.
- Quote paper
- Julia Markhovski (Author), 2024, The Feasibility of Predicting Financial Crises using Machine Learning, Munich, GRIN Verlag, https://www.grin.com/document/1453635