An application of supervised and unsupervised learning approaches to telecommunications fraud detection

Hilas, Constantinos S.; Mastorocostas, Paris As.

doi:10.1016/j.knosys.2008.03.026

Cited by 86 publications

(34 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The studies of Davey et al (1996) and Hilas and Mastorocostas (2008) (telecommunications fraud), Dorronsoro et al (1997) (credit card fraud), and Fanning and Cogger (1998), Green and Choi (1997) and Kirkos et al (2007) (financial statement fraud) all use neural network technology for detecting fraud in different contexts. Lin et al (2003) apply a fuzzy neural net, also in the domain of fraudulent financial reporting.…”

Section: Fraud Detection/prevention Literature Reviewmentioning

confidence: 99%

A Framework for Internal Fraud Risk Reduction at IT Integrating Business Processes: The IFR² Framework

Jans¹,

Lybaert²,

Vanhoof³

2009

IJDAR

View full text Add to dashboard Cite

Abstract. Fraud is a million dollar business and it is increasing every year. Both internal and external fraud present a substantial cost to our economy worldwide. A review of the academic literature learns that the academic community only addresses external fraud and how to detect this type of fraud. Little or no effort to our knowledge has been put in investigating how to prevent ánd to detect internal fraud, which we call 'internal fraud risk reduction'. Taking together the urge for research in internal fraud and the lack of it in academic literature, research to reduce internal fraud risk is pivotal. Only after having a framework in which to implement empirical research, this topic can further be investigated. In this paper we present the IFR² framework, deduced from both the academic literature and from current business practices, where the core of this framework suggests to use a data mining approach.

show abstract

Section: Fraud Detection/prevention Literature Reviewmentioning

confidence: 99%

A Framework for Internal Fraud Risk Reduction at IT Integrating Business Processes: The IFR² Framework

Jans¹,

Lybaert²,

Vanhoof³

2009

IJDAR

View full text Add to dashboard Cite

show abstract

“…In fact, a set of patterns are gathered into clusters based on similarity among each cluster. Clustering is an important technique applied in many application domains including document clustering [19]fraud detection [18], flow shop scheduling [29], machine learning [3], wireless mobile sensor networks [31], biomedical data [13], image processing [49], demand forecast [42]and financial classifications [34]. Many data clustering algorithms have been presented in the previous literatures with different approaches.…”

Section: Introductionmentioning

confidence: 99%

An Optimized K-Harmonic Means Algorithm Combined with Modified Particle Swarm Optimization and Cuckoo Search Algorithm

Bouyer

2016

Foundations of Computing and Decision Sciences

View full text Add to dashboard Cite

Abstract. Among the data clustering algorithms, k-means (KM) algorithm is one of the most popular clustering techniques due to its simplicity and efficiency. However, k-means is sensitive to initial centers and it has the local optima problem. K-harmonic-means (KHM) clustering algorithm solves the initialization problem of k-means algorithm, but it also has local optima problem. In this paper, we develop a new algorithm for solving this problem based on an improved version of particle swarm optimization (IPSO) algorithm and KHM clustering. In the proposed algorithm, IPSO is equipped with Cuckoo Search algorithm and two new concepts used in PSO in order to improve the efficiency, fast convergence and escape from local optima. IPSO updates positions of particles based on a combination of global worst, global best with personal worst and personal best to dynamically be used in each iteration of the IPSO. The experimental result on five real-world datasets and two artificial datasets confirms that this improved version is superior to k-harmonic means and regular PSO algorithm. The results of the simulation show that the new algorithm is able to create promising solutions with fast convergence, high accuracy and correctness while markedly improving the processing time.

show abstract

“…Therefore, the classification performance on the majority class would be much better than that on the minority class. Many applications that address the class imbalance problem include detection of oil spills (Kubat et al 1998), fraud detection (Fawcett and Provost 1997;Hilas and Mastorocostas 2008), credit cards (Chan et al 1999), risk management (Daskalaki et al 2006;Huang et al 2006), tornado prediction (Adrianto et al 2010), hard disk drive defect detection (Chetchotsak and Pattanapairoj 2010), and medical research (Cohen et al 2006;Mazurowski et al 2008;Ganji et al 2010;Li et al 2010). Moreover, there are several works that attempt to solve imbalanced data problems.…”

Section: Introductionmentioning

confidence: 99%

Integrating new data balancing technique with committee networks for imbalanced data: GRSOM approach

2015

View full text Add to dashboard Cite

To deal with imbalanced data in a classification problem, this paper proposes a data balancing technique to be used in conjunction with a committee network. The proposed data balancing technique is based on the concept of the growing ring self-organizing map (GRSOM) which is an unsupervised learning algorithm. GRSOM balances the data through growing new data on a well-defined ring structure, which is iteratively developed based on the winning node nearby the samples. Accordingly, the new balanced data still preserve the topology of the original data. The performance of our proposed method is evaluated using four real data sets from the UCI Machine Learning Repository and the classification performance is measured using the fivefold cross validation method. Classifiers with most common data balancing techniques, namely the Minority Over-Sampling Technique (SMOTE) and the Random under-sampling Technique (RT), are used as the baseline methods in this study. The results reveal that a committee of classifiers constructed using GRSOM performs at least as well as the baseline methods. The results also suggest that classifiers constructed using neural networks with the backpropagation algorithm are more robust than those using the support vector machine.

show abstract

An application of supervised and unsupervised learning approaches to telecommunications fraud detection

Cited by 86 publications

References 12 publications

A Framework for Internal Fraud Risk Reduction at IT Integrating Business Processes: The IFR² Framework

A Framework for Internal Fraud Risk Reduction at IT Integrating Business Processes: The IFR² Framework

An Optimized K-Harmonic Means Algorithm Combined with Modified Particle Swarm Optimization and Cuckoo Search Algorithm

Integrating new data balancing technique with committee networks for imbalanced data: GRSOM approach

Contact Info

Product

Resources

About