Page 275 - Data Science Algorithms in a Week

256 Oloruntomi Joledo, Edgar Gutierrez and Hatim Bukhari

Keywords: agent-based simulation, neural networks, consumer-to-consumer ecommerce,
peer-to-peer lending

INTRODUCTION

Organizations face an ever increasing number of challenges and threats – changes in
market, competitors, customer demands and security. To achieve organizational goals in
the midst of conflicting objectives, processes and activities need to be synchronized,
coordinated and integrated (Helal, 2008; Helal et al., 2007). Ecommerce systems are
characterized by frequent transactions from a varied customer base and consequent
reduction in order size while maintaining an element of stochasticity in demand patterns.
As a result, management faces the challenge of implementing the right strategy in the
face of competing objectives.
Peer-to-peer lending is a form of consumer-to consumer ecommerce whereby lenders
pool their resources together and lend it to borrowers at a lower rate using an online
platform without the direct mediation from financial institutions. Consumer-to-consumer
(C2C) companies face competitions from large organizations as well as from
entrepreneurs who have little to lose by embarking in the business. Customers do not
need to leave the comforts of their homes to find better deals. They can compare the
offerings of different companies online and make a hassle free change if they are not
getting value for their money. Other challenges facing C2C business models include how
to unify a group of consumers according to their needs, preferences and interaction with
each other.
Stakeholders range from providers, customers, companies and complementors (Wu
and Hisa, 2004). These stakeholders include the community, suppliers, alliance partners,
shareholders and government that form a large collection of active objects in the system
seeking to maximize their utility. With the growing popularity of C2C models, decision
making on the part of stakeholders can be difficult due to the interplaying factors and
uncertainty in customer demand. On the other hand, risks can include fidelity, payment
fraud and viruses. These characteristics make for a complex system with multi-level
abstractions and heterogeneous elements. Simulation serves as a decision support tool but
there exist limitations of individual simulation paradigms. It is in the interest of these
complex organizational environments to use knowledge of stakeholder actions and
business processes for decision-making (Joledo, 2016). These actions give rise to
nonlinear interactions that are difficult to capture using standalone simulation paradigms.
The complex interactions among different functional areas require modeling and
analyzing the system in a holistic way. There is a lack of mechanism to facilitate
systematic and quantitative analysis of the effects of users and management actions on
peer-to-peer lending system performance through the understanding of the system
behavior.