在数字化转型浪潮席卷全球的社会背景下,保险行业正以前所未有的速度向智能化、个性化服务迈进。为了精准捕捉客户需求,并有效提升旅游保险市场的渗透率,保险行业正积极探索运用高级数据分析与机器学习技术来预测客户的投保意愿。为此,...在数字化转型浪潮席卷全球的社会背景下,保险行业正以前所未有的速度向智能化、个性化服务迈进。为了精准捕捉客户需求,并有效提升旅游保险市场的渗透率,保险行业正积极探索运用高级数据分析与机器学习技术来预测客户的投保意愿。为此,我们创新性地提出了基于多模型融合策略的XGBoost-LightGBM预测模型,该模型旨在通过深度挖掘客户数据,为旅游保险产品的精准营销提供科学依据。以Kaggle平台的客户旅游保险投保情况数据集为对象进行预处理,分别构建了XGBoost与LightGBM这两个高效、灵活的梯度提升框架模型,在模型融合阶段使用采样器对模型进行参数优化,创造性地引入了AUC-RW算法确定融合权重,将两个模型的预测结果加权结合作为XGBoost-LightGBM组合模型的预测结果。最后结合准确率、F1值等评价指标,与其他算法模型进行比较分析。结果表明:结合AUC-RW算法的XGBoost-LightGBM模型相较于XGBoost、LightGBM、随机森林(RF)、支持向量机(SVM)更具有优势,预测精度更高。Amidst the global wave of digital transformation, the insurance industry is advancing towards intelligent and personalized services at an unprecedented pace. To accurately capture customer needs and effectively enhance the penetration rate of the travel insurance market, the insurance industry is actively exploring the application of advanced data analytics and machine learning techniques to predict customers’ willingness to purchase insurance. To this end, we innovatively propose an XGBoost-LightGBM prediction model based on a multi-model fusion strategy. This model aims to provide a scientific basis for precision marketing of travel insurance products by deeply mining customer data. Using the dataset of customer travel insurance purchases from the Kaggle platform as the subject of preprocessing, we have constructed two efficient and flexible gradient boosting framework models, namely XGBoost and LightGBM. During the model fusion stage, we optimize the model parameters using samplers, and creatively introduce the AUC-RW algorithm to determine the fusion weights. The prediction results of the two models are then weighted and combined to serve as the prediction outcome of the XGBoost-LightGBM ensemble model. Finally, we conduct a comparative analysis with other algorithmic models using evaluation metrics, such as accuracy and F1 score. The results indicate that the XGBoost-LightGBM model combined with the AUC-RW algorithm outperforms XGBoost, LightGBM, Random Forest (RF), and Support Vector Machine (SVM) in terms of prediction accuracy.展开更多
文摘在数字化转型浪潮席卷全球的社会背景下,保险行业正以前所未有的速度向智能化、个性化服务迈进。为了精准捕捉客户需求,并有效提升旅游保险市场的渗透率,保险行业正积极探索运用高级数据分析与机器学习技术来预测客户的投保意愿。为此,我们创新性地提出了基于多模型融合策略的XGBoost-LightGBM预测模型,该模型旨在通过深度挖掘客户数据,为旅游保险产品的精准营销提供科学依据。以Kaggle平台的客户旅游保险投保情况数据集为对象进行预处理,分别构建了XGBoost与LightGBM这两个高效、灵活的梯度提升框架模型,在模型融合阶段使用采样器对模型进行参数优化,创造性地引入了AUC-RW算法确定融合权重,将两个模型的预测结果加权结合作为XGBoost-LightGBM组合模型的预测结果。最后结合准确率、F1值等评价指标,与其他算法模型进行比较分析。结果表明:结合AUC-RW算法的XGBoost-LightGBM模型相较于XGBoost、LightGBM、随机森林(RF)、支持向量机(SVM)更具有优势,预测精度更高。Amidst the global wave of digital transformation, the insurance industry is advancing towards intelligent and personalized services at an unprecedented pace. To accurately capture customer needs and effectively enhance the penetration rate of the travel insurance market, the insurance industry is actively exploring the application of advanced data analytics and machine learning techniques to predict customers’ willingness to purchase insurance. To this end, we innovatively propose an XGBoost-LightGBM prediction model based on a multi-model fusion strategy. This model aims to provide a scientific basis for precision marketing of travel insurance products by deeply mining customer data. Using the dataset of customer travel insurance purchases from the Kaggle platform as the subject of preprocessing, we have constructed two efficient and flexible gradient boosting framework models, namely XGBoost and LightGBM. During the model fusion stage, we optimize the model parameters using samplers, and creatively introduce the AUC-RW algorithm to determine the fusion weights. The prediction results of the two models are then weighted and combined to serve as the prediction outcome of the XGBoost-LightGBM ensemble model. Finally, we conduct a comparative analysis with other algorithmic models using evaluation metrics, such as accuracy and F1 score. The results indicate that the XGBoost-LightGBM model combined with the AUC-RW algorithm outperforms XGBoost, LightGBM, Random Forest (RF), and Support Vector Machine (SVM) in terms of prediction accuracy.
