Predictive Analytics for Malware Detection in FinTech using Machine Learning Classification

No Thumbnail Available
Spelman, Fiona
Issue Date
MSc in Financial Technology
Dublin Business School
Items in eSource are protected by copyright. Previously published items are made available in accordance with the copyright policy of the publisher/copyright holder.
Cyber-attacks are a major issue in the FinTech space, and a solution is needed that can provide a fast and effective way of malware detection. This paper aims to use machine learning classification to detect malware on computers using the Microsoft Malware dataset. The research followed Cross Industry Standard Process for Data Mining (CRISP-DM) methodology and comparatively analysed Logistic Regression, Decision Trees, and Naïve Bayes models. Gaussian Naïve Bayes Classifier was the best model with a recall score of 76%. The split of the data that achieved the best result was at 70% train, 30% test. Ensemble methods were deemed unnecessary as they did not improve the recall score of the individual model. The most important features related to the size of the system on a computer, its build type, and products installed on it. It is recommended that FinTech companies use Gaussian Naïve Bayes modelling for intrusion detection systems.