Credit Risk Evaluation in Peer-to-peer Lending With Linguistic Data Transformation and Supervised Learning

Abstract

The widespread availability of various peer-to-peer lending solutions is rapidly changing the landscape of financial services. Beside the natural advantages over traditional services, a relevant problem in the domain is to correctly assess the risk associated with borrowers. In contrast to traditional financial services industries, in peerto-peer lending the unsecured nature of loans as well as the relative novelty of the platforms make the assessment of risk a difficult problem. In this article we propose to use traditional machine learning methods enhanced with fuzzy set theory based transformation of data to improve the quality of identifying loans with high likelihood of default. We assess the proposed approach on a real-life dataset from one of the largest peer-to-peer platforms in Europe. The results demonstrate that (i) traditional classification algorithms show good performance in classifying borrowers, and (ii) their performance can be improved using linguistic data transformation.

Crowdfunding: An Exploratory Study on Knowledge, Benefits and Barriers Perceived by Young Potential Entrepreneurs

“…[and] may be caused by the social pressure of conformity or by the common rationale that it is unlikely such a large group could be wrong". Since crowdfunders are not professional investors and probably do not have the appropriate skills for evaluating the risks involved in each project, "most lenders tend to follow herding behavior and consequently finance loans with high number of bidders" (Mezei 2018(Mezei , p. 1367. This herding behavior describes, according to Lee and Leem (2011, p. 495), "many social and economic situations in which an individual's decision-making is highly influenced by the decisions of others".…”

Credit Risk Classification in Peer-to-Peer Marketplaces: The Nexus of Neural Network Approach

“…Popular credit rating techniques adopted over the years include the logistic regression by Tan, Baah, Ding, Owusu-Ansah, and Agyemang (2019), artificial neural network -ANN (Byanjankar, Heikkilä , & Mezei, 2015), vector support machines (Bellotti & Crook, 2009), decision trees (Ince & Aktan, 2009), discriminant analysis (West, 2000) and nearest K-neighbors (Twala, 2010). In addition to the machine learning techniques mentioned above, random forests (Malekipirbazari & Aksakalli, 2015), linguistic interval analyzes (Mezei, Byanjankar, & Heikkilä , 2018) and survival analysis (Baesens, Setiono, Mues, & Vanthienen, 2003) are part of the recently applied credit risk assessment methods examined.…”

“…With regard to Mezei et al (2018), the use of interval-valued linguistic labels and entropy-based discretization can enhance the classification performance of traditional suppervised learning. This study proposed the use of traditional machine learning methods enhanced by data transformation based on fuzzy set theory to improve the quality of loan identification with high probability of failure.…”

When Function Meets Emotion, Change Can Happen: Societal Value Propositions and Disruptive Potential in FinTechs

“…While services of incumbent investment intermediaries seem to remain overly complex [47] and expensive [48,49], robo-advisory FTs try to respond to a growing consumer demand and are highly attractive for less privileged investors with ambitions to participate in the financial markets. Other examples of data science usecases are new authentication and access control mechanisms [42], algorithms for pattern recognition, artificial advice, ESG portfolio building and alternative risk and insurance evaluations [50,51]. Rizk, et al [52] already combine DSI with big data analytics in their general review and research agenda presenting insights that fit well to FTs as providers of DSI.…”