Detecting Insider Threats Using VADER and NAÏVE BAYS Algorithms

Nowadays larger organizations are dealing with enormous amount of emerging Security Big Data, with millions of security events being collected, processed, analyzed and responded every day, that makes it humanly impossible to detect complex zero-day cyber and insider threats. Detecting insider threats is very challenging as it deals with detection of hidden negative indicators related to the disgruntle employees with approved and authorized access to all enterprise assets. Therefore, in order to discover such complex insider threats, Natural Language Processing (NLP) & Machine Learning (ML) Algorithms can play a key role to supercharge detection and analysis of employee sentiments from textual security datasets; based on general mood, attitude, and opinions of users. This research paper analyzing and comparing two of the most common Sentiment Analysis algorithms, VADER and Naive Bayes; based on their accurate detection, real-time performance, complexity, and cost.