RESEARCH TITLE: Detecting online consumer fraud with data science techniques

Abstract: The internet provides vast amounts of information. More and more people engage in online activities through social media platforms (Facebook, Twitter, etc.) or online markets (e-bay, Amazon, etc.). Such environments enable individuals with malicious intentions to affect massive amounts of people. Authorities have the problem of finding such individuals. The amount of data available is simply too much to be analyzed by humans alone. However, the current information technology enables proficient analyses and inferences from large amounts of data, which can support authorities in their work.

Automated methods, such as machine learning (ML) or natural language processing (NLP) techniques are a viable solution to that problem. These techniques are able to meaningful analyse enormous amounts of data. NLP methods can extract grammatical or semantic information from text. For example, finding linguistic commonalities of fraudulent advertisements can be utilised to train ML classifiers, which can then categorise advertisements in being fraudulent or non-fraudulent.

It is aimed to utilise such automated methods to support authorities in their work with huge amounts of data. Gathering insights from companies and authorities from past crimes will help to establish criteria, which the automated methods will operate on. The goal of this research is to combine human knowledge and data science techniques to detect and prevent online fraud.

About Felix – click here to Read More

Graduate teaching assistant in: Applied Data Science