A Robust Unsupervised Method for Fraud Rate Estimation

Abstract

If one is interested in managing fraud, one must measure the fraud rate to be able to assess the degree of the problem and the effectiveness of the fraud management technique. This article offers a robust new method for estimating fraud rate, PRIDIT‐FRE (PRIDIT‐based Fraud Rate Estimation), developed based on PRIDIT, an unsupervised fraud detection method to assess individual claim fraud suspiciousness. PRIDIT‐FRE presents the first nonparametric unsupervised estimator of the actual rate of fraud in a population of claims, robust to the bias contained in an audited sample (arising from the quality or individual hubris of an auditor or investigator, or the natural data‐gathering process through claims adjusting). PRIDIT‐FRE exploits the internal consistency of fraud predictors and makes use of a small audited sample or an unaudited sample only. Using two insurance fraud data sets with different characteristics, we illustrate the effectiveness of PRIDIT‐FRE and examine its robustness in varying scenarios.

Just about everybody doing the business? Explaining ‘cash-for-crash’ insurance fraud in the United Kingdom

“…Traditionally such claims have been dealt with by the insurance companies themselves by simply repudiating the claims through their own internal systems of 'justice' (such private systems have been noted by other researchers such as Shearing & Stenning, 1982). This still happens with most cases and insurers have developed a variety of sophisticated detection models rooted in detailed statistical analysis, amongst others, to increase the risk of detection (Ai, Brockett, Golden, & Guillen, 2013;Artis, Ayuso, & Guillen, 2002;Caudill, Ayuso, & Guillen, 2005). However, over the last 10 years or more the insurance industry has also been 'deepening' its response to this problem by encouraging, facilitating and now paying for the police to investigate this type of fraud, amongst a number of other strategies (see Button and Brooks, 2016).…”

Credit Card Detection System Based on Ridit Approach

“…For examples, the calculation of the RIDIT score and applied Principle Component Analysis (PCA) to assign the weighted values based on the importance of the variables in insurance claim data is modified [16]. Furthermore, the RIDIT analysis to cater for ordered continuous variables is extended [22]- [24]. Newly, the categorical and continuous response variables in the detection of fraud in the microinsurance industry are mixed [15].…”

“…The summary of mean RIDIT scores for each of the response variables as a reference group is shown in Table 6. Here, R-A30 refers to A30 as the reference group; R-A31 refers to A31 as the reference group and so on [24].…”

Fraud Detection by a Multinomial Model: Separating Honesty from Unobserved Fraud

“…Insurance companies must deal with fraud perpetrated by consumers on the firm and spend money on fraud detection and monitoring. A lot of research has focused on the fraud detection efforts and the frequency of fraud, that is, assessing and ranking the fraud suspiciousness of individual claims (Ai et al, 2009(Ai et al, , 2013Artís et al, 1999;Brockett et al, 2002;Derrig and Ostaszewski, 1995;Viaene et al, 2002).…”