Data Avalanche: Harnessing for Mobile Payment Fraud Detection Using Machine Learning

Data Avalanche: Harnessing for Mobile Payment Fraud Detection Using Machine Learning

Emmanuel Awuni Kolog, Acheampong Owusu, Samuel Nii Odoi Devine, Edward Entee
DOI: 10.4018/978-1-7998-2610-1.ch002
(Individual Chapters)
No Current Special Offers


Globalizing businesses from developing countries require a thoughtful strategy and adoption of state-of-the-art technologies to meet up with the rapidly changing society. Mobile money payment service is a growing service that provides opportunities for both the formal and informal sectors in Ghana. Despite its importance, fraudsters have capitalized on the vulnerabilities of users to defraud them. In this chapter, the authors have reviewed existing data mining techniques for exploring the detection of mobile payment fraud. With this technique, a hybrid-based machine learning framework for mobile money fraud detection is proposed. With the use of the machine learning technique, an avalanche of fraud-related cases is leveraged, as a corpus, for fraud detection. The implementation of the framework hinges on the formulation of policies and regulations that will guide the adoption and enforcement by Telcos and governmental agencies with oversight responsibilities in the telecommunication space. The authors, therefore, envision the implementation of the proposed framework by practitioners.
Chapter Preview


Information and communication technologies (ICT) have revolutionized the operations of business enterprises enabling them to embrace more flexible yet efficient approaches to improve customer intimacy. The pervasive impact of ICT can be seen in the developed economies where the transformation has been rapid, with frequent technological innovations to meet societal changes. Conversely, the pace of digital innovations in the developing economies, especially Africa, has been slow over the period (UNCTAD Report, 2018) with many of the businesses relying on the west for technological innovations. This can be attributed to the low level of education on the continent. Nevertheless, the digital revolution has not only brought about efficiency in human activities but has heightened competition in all aspects of business activities. These competitions are not only among businesses wanting to survive but also among individuals competing for superiority. People, for example, are engrossed with technology and inherently desire for the latest and fastest available technologies, such as high-speed mobile phones. The urge for these advances in technological innovations essentially stems from the desire to compete and survive in an increasingly competitive business environment (Obrdovic, 2016). Essentially, while trying to survive in the business ecosystem, the innovative paradigm is geared towards the integration of ICT into automating business processes.

Every serious organization seeks to commit resources into technology innovations in order to survive (Baron & Spulber, 2017). Disruptiveness is the result of many technological breakthroughs. Disruptive innovation is one that generates or drives the formation of new markets and value networks which ultimately leads to the disruption of current markets and the entire value chain from the organizational perspective to their products and related services (Rahman et al., 2018). Presently, to survive and compete favourably through innovative approaches, the paradigm leans towards technology. Accordingly, many companies have invested heavily in ICT infrastructure to stay in business by improving customer intimacy. These businesses must keep challenging the technological status quo at any given time to rapidly create value for customers thereby motivating a disruptive technology. Notably, while the current generation is seamlessly focusing on improving human interactions and communications through technology, there exists the side effect of technology usage. Sadly, as many businesses and individuals are capitalizing on harnessing the potential of the internet to improve their business decisions, others use the internet to commit crimes. In Africa, these crimes, often through social engineering, are committed by fraudsters who take advantage of people’s ignorance and greed. In Section 4, these authors have delved into some social engineering strategies that fraudsters use in defrauding mobile money users in Ghana.

Globalizing businesses from developing countries require a thoughtful strategy and adoption of state-of-the-art technologies in order to meet up with a rapidly changing global society. To be competitive in the business environment, a profound understanding of customers’ behaviours towards a brand, services and/or products is required. The avalanche of unstructured and structured data, from the web or operational databases, can be harnessed to understand customers’ behaviour through data mining with machine learning techniques (Femima & Sudheep (2015). However, many businesses in the developing countries have not capitalized on the potentials of data mining, such as classification, forecasting and clustering, to inform business decisions. This partly informs the reason for this chapter. This situation has been largely attributed to the level of ignorance and lack of knowledge exhibited by corporate managers and related portfolios in the developing country context.

As we are inundated with an avalanche of data, this chapter presents the impact of data mining on business decisions. This chapter is grounded on a review of machine learning approaches with linkage to the developing country context, especially in Africa. Based on the review, the authors have proposed a data mining with a machine learning technique to detect fraud in mobile money activities in Ghana. The technique, also presented as a framework, is to provide real-time detection of fraud through mobile communications. The authors envision that this proposed framework will be adopted by the telecommunication companies to curb the menace on mobile money fraud.

Key Terms in this Chapter

Classification: This a supervised learning where the input data is tagged with an output data. The goal of classification is to predict the output data based on the input data.

Data: Data is raw fact that can be processed for information. Examples of data include texts, images, and videos.

Data Avalanche: Is a big data that can computationally be analysed to reveal patterns, trends, and associations.

Machine Learning: Machine learning focuses on the development of computer programs that can access data and use it to learn for themselves.

Data Mining: Data mining is a process that focuses on identifying correlations, relevant patterns, and trends, meticulously selected from massive datasets resident in some repository.

Clustering: Clustering is an unsupervised learning technique where the input is not tagged. Clustering identify patterns in the data based on datapoint similarities.

Complete Chapter List

Search this Book: