To what extent can transactions be identified as a fraud?

Research Question Context Fraud monitoring and prevention are the most challenging and costly financial businesses. For example, in 2018, $24.26 Billion was lost due to payment card fraud. Banks and financial houses try to reduce fraud by investing money in software development. The United States leads as the most credit fraud-prone country, with 38.6% of reported card fraud losses in 2018. (Shiftprocessing, 2022) Analyzing the features of the transactions using machine learning like the logistic regression model could statistically significantly identify fraudulent transactions, and the study results could be used as proof of concept to develop applications in the future for fraud monitoring and prevention. Machine learning can be a powerful and influential tool in one of the most challenging and restricted sectors and will drive to increase the trust for safe transactions and more financial revenue. Research Question The Financial houses and Banks are still looking for tools to monitor and prevent fraud, and at the same time looking to measure the accuracy, efficiency, and effectiveness that could be summarized in one research question: "To what extent can transactions be identified as a fraud?" Justification The research question covers the financial houses’ and banks' actual needs and determines if the transactions could identify as fraud. This research question covers both requirements: the ability and accuracy of the same question. The research results that will answer the research question will provide the details that will help the decision-maker use the research model as a proof of concept. The research question presents the opportunity for the data analytics and data researcher to compare the results, like comparing the predicted data with the test data to define if the model can identify the fraudulent transactions and to what extent. Hypothesis The research hypothesis will be: "Fraudulent transactions can statistically significantly be identified from the provided dataset." The research will evaluate if it can statistically significantly identify fraud transactions. The evidence will be collected to confirm or reject the hypothesis from the logistic regression model as one of the machine learning models. The model evaluation will determine if the thesis can be validated or rejected.The research question covers the financial houses' and banks' actual needs and determines if the transactions could identify as fraud. This research question covers both requirements: the ability and accuracy of the same question. The research results that will answer the research question will provide the details that will help the decision-maker use the research model as a proof of concept. The research question presents the opportunity for the data analytics and data researcher to compare the results, like comparing the predicted data with the test data to define if the model can identify the fraudulent transactions and to what extent