CS BSc Thesis Presentation 2 December 2022
Plats: Online via: https://lu-se.zoom.us/j/67564985598?pwd=anJzeVU5RVJwZ2xZdDRrdWFnQzZJZz09
Kontakt: birger [dot] swahn [at] cs [dot] lth [dot] se
Spara händelsen till din kalender
One Computer Science BSc thesis to be presented on 2 December
Friday, 2 December there will be a bachelor thesis presentation in Computer Science at Lund University, Faculty of Engineering.
N.B. The presentation will only take place in Zoom (https://lu-se.zoom.us/j/67564985598?pwd=anJzeVU5RVJwZ2xZdDRrdWFnQzZJZz09).
Note to potential opponents: Register as an opponent to the presentation of your choice by sending an email to the examiner for that presentation (email@example.com). Do not forget to specify the presentation you register for! Note that the number of opponents may be limited (often to two), so you might be forced to choose another presentation if you register too late. Registrations are individual, just as the oppositions are! More instructions are found on this page.
09:15-10:00 Online in Zoom
Presenters: Daniel Amir, Max Söderlind
Title: Analysis of different machine learning approaches for financial transactional fraud detection
Examiner: Marcus Klang
Supervisor: Dennis Medved (Lund University)
Credit card fraud is a societal problem affecting millions of people every year resulting in losses of billions of dollars. Detecting fraudulent transactions is difficult for the human eye and through the rise of technology, there are now other, more complex and accurate solutions to this problem using machine learning and artificial intelligence. The aim of this thesis is to assess how different machine learning algorithms can be used to accurately predict fraudulent credit card transactions. In this study, 3 different machine learning models were implemented. One random forest model, one neural network model, and an ensemble model, based on a random forest. Moreover, hyperparameter tuning algorithms were used to maximize the accuracy of detecting fraudulent transactions. The models were tested with a dataset that consisted of 284,807 transactions with 492 frauds. The finding of this work was that the ensemble model outperformed the neural network model and the random forest model in detecting fraudulent transactions with an accuracy of 93.3%, but performed worst in detecting non-fraudulent transactions with an accuracy of 96.0%. Furthermore, there were small differences in accuracy between the models and to further improve the accuracy of detecting fraudulent transactions, more complex machine learning models should be implemented.