Thesis Abstract

Abstract
This thesis explores the application of machine learning techniques for enhancing fraud detection in the context of online banking. With the increasing digitalization of financial services, the risk of fraudulent activities has also escalated, posing a significant threat to both financial institutions and customers. Traditional rule-based fraud detection systems are often limited in their ability to adapt to evolving fraud patterns and can result in high false positive rates. Machine learning algorithms offer a promising solution by enabling automated fraud detection through the analysis of large volumes of transaction data. The research begins with an introduction to the challenges of fraud detection in online banking, highlighting the importance of developing more advanced and efficient detection methods. The background of the study provides a comprehensive overview of the evolution of fraud detection in the banking sector and the emergence of machine learning as a powerful tool in this domain. The problem statement emphasizes the need for more sophisticated fraud detection techniques to combat the growing threat of online banking fraud. The objectives of the study include evaluating the effectiveness of various machine learning algorithms in detecting fraudulent activities, enhancing the accuracy of fraud detection models, and minimizing false positive rates. The limitations of the study are also identified, including constraints related to data availability, algorithm performance, and model interpretability. The scope of the study defines the boundaries within which the research is conducted, focusing specifically on the application of machine learning for fraud detection in online banking. The significance of the study lies in its potential to contribute to the development of more robust and adaptive fraud detection systems that can effectively mitigate risks associated with online banking fraud. The structure of the thesis outlines the organization of the research, including the chapters dedicated to literature review, research methodology, discussion of findings, and conclusion. The literature review examines existing research on fraud detection in online banking, highlighting the strengths and limitations of different machine learning approaches. Key topics covered include data preprocessing techniques, feature selection methods, model evaluation metrics, and the comparison of various machine learning algorithms for fraud detection. The research methodology section details the data collection process, feature engineering techniques, model selection criteria, and evaluation methodology employed in the study. Various machine learning algorithms, such as logistic regression, random forest, support vector machines, and neural networks, are implemented and compared based on their performance metrics, including accuracy, precision, recall, and F1 score. The discussion of findings presents the results of the experiments conducted, highlighting the strengths and weaknesses of different machine learning algorithms in detecting fraudulent activities in online banking transactions. The analysis includes insights into the factors influencing model performance, such as data quality, feature importance, and algorithm complexity. In conclusion, the study underscores the potential of machine learning applications for enhancing fraud detection in online banking and provides recommendations for further research and practical implementation. By leveraging advanced machine learning techniques, financial institutions can strengthen their defenses against online banking fraud and safeguard the interests of their customers. Keywords Machine learning, Fraud detection, Online banking, Financial services, Data analysis, Algorithm, Model evaluation, Risk management.

Thesis Overview