Abstract— This work is on money laundering discovery in monetary organizations with a cascaded artificial neural network and k-medoids approach. The overbearing of averting cash laundering wrongdoings is not because the launderers takes unjustified benefit of the economy for illegal financial achievement but also since cash laundering is a misconduct that is typically associated to criminalities like fraud, Illegal trafficking in drugs, abduction, corruption, arms trafficking, terrorism, among others. This research used a cascade model of K-medoids and Artificial neural network (ANN) to design a robust detection model for money laundering. K-medoids method is used to cluster the dataset into two clusters, its output is filtered, and then passed into ANN for training and testing purposes and classified into doubtful and un-doubtful transactions. Other algorithms like Support Vector Machine (SVM) and ANN were used separately on the dataset for detection purposes. Their performances were matched with the Cascade K-medoid-ANN. Results shows that ANN has an accuracy of 80.3%. SVM has an accuracy of 70.6% and Cascade K-medoid-ANN has an accuracy of 74.5%. Other metrics used includes; Sensitivity, Specificity, and Precision. Cascade K-medoid-ANN outperformed other algorithms used by correctly identifying doubtful transactions as doubtful and un-doubtful transactions to be un-doubtful. It is recommended that all monetary organizations work together to offer a warehouse of cash laundering dataset for tackling menace and other bank frauds. Index Terms— Artificial Neural Network, Financial organization, K-Medoids, Money laundering I. INTRODUCTION RTIFICIAL Artificial neural network (ANN) encompasses of elements termed processed neurons. Artificial neuron tries to replicate the structure of natural neuron and conduct. A neuron contains of single yield (synapse through axon) and single input or dendrites. The neuron's activation defines its feature. A neuron is a component that procedures information that is vital for the operation of ANN. It is a parallel spread processor containing of easy handling components with a disposition which stock and make available for empirical knowledge usage. It looks like the brain: the network obtains Manuscript received December 28, 2019; revised February 3, 2020. J. K. Alhassan is with the Federal University of Technology, Department of Computer Science, Niger State, Minna, Nigeria, Phone: +2348035961620; (e-mail: jkalhassan@futminna.edu.ng) N. E. Yusuf was with the Federal University of Technology, Department of Computer Science, Niger State, Minna, Nigeria, (e-mail: nimah93@gmail.com). information from its setting vial a technique of learning and inter-neuron connection powers, denoted as synaptic weights, they are used to stock the knowledge gotten. ANN have been useful in diverse areas such as monetary market predicting, credit card fraud discovery and risk assessment. The k-medoids is an algorithm which is used for clustering which partitions (dividing the dataset to clusters) and try to decrease the space amid points considered to be in a group and a point allocated the centre of that group. K- medoids selects data plugs as centres (exemplars or medoids) random distances can be used. These k medoids denotes diverse essential facets of the scope N data set that are separated to exhaustive clusters and mutually exclusive around k medoids, wherever a medoid is the entity of the cluster which the amount of spaces to all other cluster is negligible. This method is robust in contradiction of outliers, noise or barely disseminated information as a result of medoids [1]. All over the world, there are huge volumes of monetary dealings vial diverse means each minute. These comprise of fraudulent transactions by criminals that have initiated damage to monetary organizations or clienteles, or cause reputational injury, that is tough to overhaul [2]. Money laundering is one of such fraudulent transaction, which is a way of hiding the source of moneys illicitly gotten and accumulated after a while [3]. By cash laundering, criminals attempt to opaque the real source of cash acquired from unlawful action. Cash laundering is a world-wide problematic issue with profuse damaging effects on society. The influence remains extra overwhelming in the emerging financial prudence with feeble monetary controlling scheme and starting where the moneys are relocated to advanced countries to obtain glamorous extravagance substances. In specific, laundering of cash offers a way of obtaining proceeds for the offenders, providing free money which may be for financing additional criminal actions. Furthermore, cash laundering can severely undermine sureness in monetary structures and financial organizations, and cause harm to local economies. It is perceived that cash laundering takes countless consequence on the Nigerian budget [4]. Notwithstanding the policies and laws passed, cash laundering with additional monetary and fiscal crimes flourish in the Nigeria [5]. This research is on a cascaded approach of ANN and K- medoid for noticing cash laundry in Nigeria by means of unidentified transaction data from manifold banking sources. The remains part of this work is organized as A Cascaded Artificial Neural Network and K-Medoids Method for Money Laundering Detection in Financial Organizations J. K. Alhassan, Member, IAENG, and N. E. Yusuf A Proceedings of the World Congress on Engineering 2021 WCE 2021, July 7-9, 2021, London, U.K. ISBN: 978-988-14049-2-3 ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online) WCE 2021