File Download
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.3390/risks8010019
- Scopus: eid_2-s2.0-85079905889
- WOS: WOS:000524496300028
- Find via
Supplementary
- Citations:
- Appears in Collections:
Article: Delta Boosting Implementation of Negative Binomial Regression in Actuarial Pricing
Title | Delta Boosting Implementation of Negative Binomial Regression in Actuarial Pricing |
---|---|
Authors | |
Keywords | Boosting trees Gradient boosting Insurance Machine learning Negative binomial Predictive modeling |
Issue Date | 2020 |
Publisher | MDPI AG. The Journal's web site is located at http://www.mdpi.com/journal/risks |
Citation | Risks, 2020, v. 8 n. 1, p. article no. 19 How to Cite? |
Abstract | This study proposes an efficacious approach to analyze the over-dispersed insurance frequency data as it is imperative for the insurers to have decisive informative insights for precisely underwriting and pricing insurance products, retaining existing customer base and gaining an edge in the highly competitive retail insurance market. The delta boosting implementation of the negative binomial regression, both by one-parameter estimation and a novel two-parameter estimation, was tested on the empirical data. Accurate parameter estimation of the negative binomial regression is complicated with considerations of incomplete insurance exposures, negative convexity, and co-linearity. The issues mainly originate from the unique nature of insurance operations and the adoption of distribution outside the exponential family. We studied how the issues could significantly impact the quality of estimation. In addition to a novel approach to simultaneously estimate two parameters in regression through boosting, we further enrich the study by proposing an alteration of the base algorithm to address the problems. The algorithm was able to withstand the competition against popular regression methodologies in a real-life dataset. Common diagnostics were applied to compare the performance of the relevant candidates, leading to our conclusion to move from light-tail Poisson to negative binomial for over-dispersed data, from generalized linear model (GLM) to boosting for non-linear and interaction patterns, from one-parameter to two-parameter estimation to reflect more closely the reality. |
Persistent Identifier | http://hdl.handle.net/10722/287891 |
ISSN | 2023 Impact Factor: 2.0 2023 SCImago Journal Rankings: 0.403 |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | LEE, SCK | - |
dc.date.accessioned | 2020-10-05T12:04:45Z | - |
dc.date.available | 2020-10-05T12:04:45Z | - |
dc.date.issued | 2020 | - |
dc.identifier.citation | Risks, 2020, v. 8 n. 1, p. article no. 19 | - |
dc.identifier.issn | 2227-9091 | - |
dc.identifier.uri | http://hdl.handle.net/10722/287891 | - |
dc.description.abstract | This study proposes an efficacious approach to analyze the over-dispersed insurance frequency data as it is imperative for the insurers to have decisive informative insights for precisely underwriting and pricing insurance products, retaining existing customer base and gaining an edge in the highly competitive retail insurance market. The delta boosting implementation of the negative binomial regression, both by one-parameter estimation and a novel two-parameter estimation, was tested on the empirical data. Accurate parameter estimation of the negative binomial regression is complicated with considerations of incomplete insurance exposures, negative convexity, and co-linearity. The issues mainly originate from the unique nature of insurance operations and the adoption of distribution outside the exponential family. We studied how the issues could significantly impact the quality of estimation. In addition to a novel approach to simultaneously estimate two parameters in regression through boosting, we further enrich the study by proposing an alteration of the base algorithm to address the problems. The algorithm was able to withstand the competition against popular regression methodologies in a real-life dataset. Common diagnostics were applied to compare the performance of the relevant candidates, leading to our conclusion to move from light-tail Poisson to negative binomial for over-dispersed data, from generalized linear model (GLM) to boosting for non-linear and interaction patterns, from one-parameter to two-parameter estimation to reflect more closely the reality. | - |
dc.language | eng | - |
dc.publisher | MDPI AG. The Journal's web site is located at http://www.mdpi.com/journal/risks | - |
dc.relation.ispartof | Risks | - |
dc.rights | This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. | - |
dc.subject | Boosting trees | - |
dc.subject | Gradient boosting | - |
dc.subject | Insurance | - |
dc.subject | Machine learning | - |
dc.subject | Negative binomial | - |
dc.subject | Predictive modeling | - |
dc.title | Delta Boosting Implementation of Negative Binomial Regression in Actuarial Pricing | - |
dc.type | Article | - |
dc.description.nature | published_or_final_version | - |
dc.identifier.doi | 10.3390/risks8010019 | - |
dc.identifier.scopus | eid_2-s2.0-85079905889 | - |
dc.identifier.hkuros | 314863 | - |
dc.identifier.volume | 8 | - |
dc.identifier.issue | 1 | - |
dc.identifier.spage | article no. 19 | - |
dc.identifier.epage | article no. 19 | - |
dc.identifier.isi | WOS:000524496300028 | - |
dc.publisher.place | Switzerland | - |
dc.identifier.issnl | 2227-9091 | - |