A recommendation system for car insurance
The authors construct a recommendation system for car insurance. The system combines the XGBoost algorithm and the \textit{a priori} algorithm to choose which customer should be recommended and which cover to recommend, respectively. XGBoost algorithm is based on gradient boosting method -- a sequential ensemble method. The principle of boosting is to combine weak learners (e. g. decision trees for gradient tree boosting) trained in sequence to build a strong learner. In gradient tree boosting, each decision tree attempts to correct errors made by the previous tree. At each iteration, a decision tree is fitted to residual error. The \textit{a priori} algorithm was introduced in order to find association rules in a dataset (e.g. a collection of items bought together). The tests show that the approach outperforms standard results for similar up-selling campaigns.
- scientific article; zbMATH DE number 7416537
- Recommendation systems: A probabilistic analysis
- Constructing a hybrid recommender system
- Improving automobile insurance claims frequency prediction with telematics car driving data
- A survey of personalized treatment models for pricing strategies in insurance
This page was built for publication: A recommendation system for car insurance
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2219620)
