Synthetic Dataset of Driver Telematics

合成驾驶员远程信息处理数据集是由康涅狄格大学数学系创建的，旨在通过模拟真实保险数据集的特性，提供一个用于研究使用基于使用保险（UBI）的风险评估模型的资源。该数据集包含100,000条政策记录，涵盖了驾驶员的索赔经验以及相关的传统风险变量和远程信息处理相关变量。数据集的创建过程分为三个阶段，使用机器学习算法模拟索赔数量和总索赔金额，并通过扩展的SMOTE算法生成特征空间。该数据集主要用于学术研究，以校准和测试精算和风险评估模型，同时也适用于市场研究，如保险公司进入UBI市场的策略分析。

The Synthetic Driver Telematics Dataset was created by the Department of Mathematics, University of Connecticut, aiming to provide a resource for researching Usage-Based Insurance (UBI)-based risk assessment models by simulating the characteristics of real insurance datasets. This dataset contains 100,000 policy records, covering drivers' claims experience, as well as relevant traditional risk variables and telematics-related variables. The dataset was developed in three stages: machine learning algorithms were used to simulate claim counts and total claim amounts, and the extended SMOTE algorithm was applied to generate the feature space. Primarily intended for academic research to calibrate and test actuarial and risk assessment models, this dataset is also applicable to market research such as strategic analysis for insurance companies entering the UBI market.