● Analyzed credit card fraud data comprising 275,663 records and 30 attributes, employing data cleaning, preprocessing, and visualization techniques for comprehensive insights.
● Strategically partitioned data into training, validation, and testing sets, addressing missing values through KNN imputation and eliminating duplicates.
● Implemented logistic regression, yielding a remarkable accuracy of 92.34%, and conducted feature selection based on independent variable importance. Refined model performance through meticulous tuning, ensuring robust accuracy to enhance fraud detection capabilities and mitigate risks effectively.