Deep Learning for Cyber Risk Management in Financial Services: A Case Study of Data Breach Prediction
Archana Pattabhi · International Journal of Emerging Research in Engineering and Technology · 2020
Amidst a rapidly growing and highly computerized environment, the financial services sector remains one of the most vulnerable to cyber threats and data breaches that have high operational and image risk implications. This paper aims to examine how deep learning can be used to identify factors that can be used in predicting data breaches, especially within the financial sector. With data from past security breaches and their metadata, the suggested deep learning model shall use LSTM and CNN networks in a multidimensional data set containing IT infrastructure data and records of user behaviour and threat intelligence feeds from third parties. We then determine the current cyber risk environment and explain why conventional risk management approaches are insufficient. We also explain how we managed the data and feature collection, derived the features from raw data, designed the architectures of the models, and evaluated the results. The performance of the proposed hybrid deep learning model has been tested on a benchmark dataset Contemplating real-world cyber incidents and compared to the results of classic machine learning methods like the Random Forests and Support Vector Machines (SVMs) where we have the numbers of 12% average of F1-score. This research is valuable because it proposes a model for measuring future cyber threats for SOC in the financial industry. Therefore, a proactive, artificial intelligence cybersecurity solution can have a prospective impact in lessening the blur incidence rate and enhancing compliance with the legislation