INVESTIGATING THE IMPACT OF IMBALANCED DATA ON CREDIT CARD FRAUD DETECTION MODELS

Abstract:

Credit card fraud has become a pervasive problem in the financial industry, resulting in substantial financial losses for both individuals and businesses. To mitigate this issue, various machine learning models have been developed for credit card fraud detection. However, the presence of imbalanced data, where the number of non-fraudulent transactions significantly outweighs the number of fraudulent transactions, poses a significant challenge to the effectiveness of such models.

This research aims to investigate the impact of imbalanced data on credit card fraud detection models. The study utilizes a real-world credit card transaction dataset, consisting of both fraudulent and non-fraudulent transactions, to evaluate the performance of different machine learning algorithms under imbalanced conditions.

First, the imbalanced dataset is preprocessed to address data quality issues, handle missing values, and normalize the features. Next, various techniques for handling class imbalance are implemented, such as oversampling the minority class (fraudulent transactions) using Synthetic Minority Over-sampling Technique (SMOTE), undersampling the majority class, and using hybrid approaches like SMOTE combined with Edited Nearest Neighbors (SMOTE-ENN).

Several popular machine learning algorithms, including logistic regression, random forest, support vector machines, and neural networks, are trained and evaluated on the imbalanced dataset using appropriate evaluation metrics such as accuracy, precision, recall, and F1-score. The performance of these models is then compared to assess their effectiveness in detecting credit card fraud.

Furthermore, the impact of different evaluation metrics on model performance is explored to determine the most suitable metric for imbalanced data scenarios. Additionally, the study investigates the trade-off between model performance and computational efficiency, as imbalanced data may lead to increased computational requirements.

The findings of this research will provide valuable insights into the challenges posed by imbalanced data on credit card fraud detection models. The results will aid in identifying effective strategies for handling class imbalance and improving the overall performance of credit card fraud detection systems. Ultimately, this research aims to contribute to the development of more robust and accurate models that can effectively detect fraudulent credit card transactions, thereby enhancing the security and trust in financial transactions.

INVESTIGATING THE IMPACT OF IMBALANCED DATA ON CREDIT CARD FRAUD DETECTION MODELS. GET MORE MASTERS COMPUTER SCIENCE 

Sharing is caring!