Enhanced gradient boosting for zero-inflated insurance claims and comparative analysis of CatBoost, XGBoost, and LightGBM

IF 1.6 3区经济学 Q3 MATHEMATICS, INTERDISCIPLINARY APPLICATIONS

Scandinavian Actuarial Journal Pub Date : 2024-06-10 DOI:10.1080/03461238.2024.2365390

Banghee So

引用次数: 0

Abstract

The property and casualty (P&C) insurance industry faces challenges in developing claim predictive models due to the highly right-skewed distribution of positive claims with excess zeros. To addres...

查看原文本刊更多论文

针对零膨胀保险理赔的增强梯度提升以及 CatBoost、XGBoost 和 LightGBM 的对比分析

由于正赔率与过多零赔率的高度右偏分布，财产与意外（P&C）保险行业在开发赔率预测模型方面面临挑战。为了解决这个问题，我们开发了一种新的理赔预测模型。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Scandinavian Actuarial Journal MATHEMATICS, INTERDISCIPLINARY APPLICATIONS-STATISTICS & PROBABILITY

CiteScore

3.30

自引率

11.10%

发文量

审稿时长

>12 weeks

期刊介绍： Scandinavian Actuarial Journal is a journal for actuarial sciences that deals, in theory and application, with mathematical methods for insurance and related matters. The bounds of actuarial mathematics are determined by the area of application rather than by uniformity of methods and techniques. Therefore, a paper of interest to Scandinavian Actuarial Journal may have its theoretical basis in probability theory, statistics, operations research, numerical analysis, computer science, demography, mathematical economics, or any other area of applied mathematics; the main criterion is that the paper should be of specific relevance to actuarial applications.