Sequence Embeddings Help Detect Insurance Fraud

Fursov, Ivan; Kovtun, Elizaveta; Rivera-Castro, Rodrigo; Zaytsev, Alexey; Khasyanov, Rasul; Spindler, Martin; Burnaev, Evgeny

doi:10.1109/access.2022.3149480

Cited by 8 publications

(3 citation statements)

References 42 publications

(35 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In recent years, several studies have examined the effects of insurtech on various aspects of insurance. Claims management, which is one of the fundamental functions of insurance, has emerged as a significant area of research (Fursov et al 2022). Claims management means collecting, recording, analyzing, tracking, and settling claims (Farbmacher et al 2022).…”

Section: Literature Reviewmentioning

confidence: 99%

Insurtech in Europe: identifying the top investment priorities for driving innovation

Eti,

Dinçer,

Meral

et al. 2024

Financ Innov

View full text Add to dashboard Cite

The purpose of this study is to determine the essential indicators to improve insurtech systems and select the most critical alternative to increase insurtech-based investments in European countries. A novel fuzzy decision-making model is generated by integrating entropy and additive ratio assessment (ARAS) techniques with spherical fuzzy sets. First, the indicators are weighted using spherical fuzzy entropy. Then, the alternatives are ranked using spherical fuzzy ARAS. The alternatives are also ranked with the spherical fuzzy technique for order of preference by similarity to the ideal solution methodology. The main contribution of this study is that it would help investors to take the right actions to increase the performance of insurtech investments without incurring high costs. Another important novelty is that a new fuzzy decision-making model is proposed to solve this problem. The results of the two models are quite similar, proving the validity and coherency of the findings. It is found that pricing is the most critical factor that affects the performance of insurtech investments. Insurtech companies are required to make accurate pricing by conducting risk analyses to increase their profits and minimize their risks. Additionally, according to the ranking results, big data are the most appropriate way to improve the performance of insurtech investments in Europe. Big data analytics helps companies learn more about the behavior of their customers. By analyzing data about their customers’ past transactions, companies can provide more convenient services to them. This would increase customer satisfaction and enable companies to achieve long-term customer loyalty.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Insurtech in Europe: identifying the top investment priorities for driving innovation

Eti,

Dinçer,

Meral

et al. 2024

Financ Innov

View full text Add to dashboard Cite

show abstract

“…This research conclude that Decision Tree gives the highest accuracy of 79% as compared to the other techniques. Further research [10] mentioned about 10 percent of the losses incurred by the insurance industry are estimated to come from fraud claim. This research processed 3.3 million health care bill data .…”

Section: Introductionmentioning

confidence: 99%

Comparative Analysis Performance of Naïve Bayes and K-NN Using Confusion Matrix and AUC To Predict Insurance Fraud

Triyono

Ginting²

2022

mib

View full text Add to dashboard Cite

Based on claim submission data from year 2019 to 2021 can be seen that the percentage of claims in one province is much higher than other provinces. During that period, the percentage of claims in that province reached 22% while the highest percentage in other provinces was only 6%. It is suspected that there has been a claim fraud in the province. The fraud allegedly started when customer submits a policy issuance for the elderly insured with a low sum insured so that the premium is also low. The insured's health condition at that time may not be good but it is not explained in the insurance application letter. To increase the sum insured, the policy is usually added with additional coverage. Fraud claim creates big loss for insurance company since the company has to pay the claim that they should not pay. Insurance company need to have a mechanism to avoid the fraud claim. From this research, it is expected to find the best methodology to be able to predict the potential of insurance claim fraud early when customers apply for policy issuance so that additional checks can be carried out for suspected submissions. The initial data available for this research is 14,778 claim records with attributes are : the date of claim submission, policy effective date, sum assured, type of claim, cause of claim, province and fraud. In order to get the best methodology on the accuracy and performance aspect to fulfill the expectation, two methodologies (Naïve Bayes and K-NN) are compared. Both Naive Bayes and K-NN methods are used with a comparison of training data and testing data is 80:20. Several combinations were performed for each of these methods. By using Confusion Matrix and AUC to measure the accuracy and performance of the two methods, it can be concluded that the best one is Naive Bayes with accuracy is 90% and AUC is 0.761. The attributes used are province, sum assured, additional coverage and the insured is the policy holder.

show abstract

“…This more detailed signal data from systems closer to the encounter identifies anomalies within and across the modeled entities. Fursov (2022) demonstrates an approach to transform relational claims data into a graph, embedding object descriptions in vectors with the same dimensionality as the graph nodes. This enables Neural Network approaches for analysis that outperform traditional machine learning approaches.…”

Section: Data Modelingmentioning

confidence: 99%

Data-driven models, techniques, and design principles for combatting healthcare fraud

Thornton

View full text Add to dashboard Cite

In the U.S., approximately $700 billion of the $2.7 trillion spent on healthcare is linked to fraud, waste, and abuse (Kelley, 2009). This presents a significant challenge for healthcare payers, including governments, insurers, and businesses, as they navigate fraudulent activities from dishonest practitioners, sophisticated criminal networks, and even wellintentioned providers who inadvertently submit incorrect billing for legitimate services. Government-run programs are particularly vulnerable to fraud, given the challenges in excluding problematic providers compared to private networks.The system's complexity, diversity of actors, and sparsity of labeled data make applying data analysis methods used in other sectors challenging. However, with careful engineering and ongoing adjustments, data analysis techniques such as outlier detection can support programs in controlling escalating costs and maintaining financial stability. This thesis adopts Hevner's (2004) research methodology to guide the creation, assessment, and refinement of a healthcare fraud detection framework and recommended design principles for fraud detection in other similarly complex environments. The thesis provides the following significant contributions to the field:1. A formal literature review of the field of fraud detection in Medicaid. Chapters 3 and 4 provide formal reviews of the available literature on healthcare fraud. Chapter 3 focuses on defining the types of fraud found in healthcare. Chapter 4 reviews fraud detection techniques in literature across healthcare and other industries. Chapter 5 focuses on literature covering fraud detection methodologies utilized explicitly in healthcare. A multidimensional data model and analysis techniques for fraud detection in healthcare.Chapter 5 applies Hevner et al. ( 2004) to help develop a framework for fraud detection in Medicaid that provides specific data models and techniques that identify the most prevalent fraud schemes. Based on the environment and knowledge base analysis, a multidimensional schema based on Medicaid data and a set of multidimensional models and techniques to detect fraud in large sets of claim transactions are presented. These artifacts are evaluated through functional testing against known fraud schemes. This chapter contributes a set of multidimensional data models and analysis techniques that can be used to detect the most prevalent known fraud types. 3.A framework for deploying outlier-based fraud detection methods in healthcare. Chapter 6 proposes and evaluates methods for applying outlier detection to healthcare fraud based on literature review, comparative research, direct application on healthcare claims data, and known fraudulent cases. Based on a multi-dimensional data model developed for Medicaid claim data (Thornton et al., 2013), a method for outlier-based fraud detection is presented and evaluated using Medicaid dental claims, providers, and patients in an actual US state Medicaid program. Design principles for fraud detection in complex systems.Based on...

show abstract

Sequence Embeddings Help Detect Insurance Fraud

Cited by 8 publications

References 42 publications

Insurtech in Europe: identifying the top investment priorities for driving innovation

Insurtech in Europe: identifying the top investment priorities for driving innovation

Comparative Analysis Performance of Naïve Bayes and K-NN Using Confusion Matrix and AUC To Predict Insurance Fraud

Data-driven models, techniques, and design principles for combatting healthcare fraud

Contact Info

Product

Resources

About