Nonlife Insurance Risk Classification Using Categorical Embedding

In nonlife insurance risk classification, many rating factors are naturally categorical and often the categorical variables have a large number of levels. The high cardinality of categorical rating variables presents challenges in actuarial applications of both traditional generalized linear models and modern deep learning methods. To address these challenges, we introduce the method of categorical embedding to the practice of risk classification. The method is inspired by the neural network language models for learning text data and maps a categorical variable into a real-valued representation in the Euclidean space. Using a property insurance claims data set, we demonstrate the use of categorical embedding in three applications. The first shows how embeddings are used to construct rating classes and calculate rating relativities for a single insurance risk. The second concerns predictive modeling for multivariate insurance risks and emphasizes the effects of dependence on tail risks. The third focuses on pricing new products where transfer learning is used to gather knowledge from existing products