Intelligent Insurance Actuarial Model Under Machine Learning and Data Mining

Traditional actuarial methods in insurance are inadequate and have low computational accuracy when dealing with massive amounts of information in the modern insurance industry. This article intended to screen out features that play an important role in insurance actuarial analysis by applying information gain, mutual information, and other means, and screen input variables to avoid dimensional disasters. In order to improve the learning speed of intelligent insurance actuarial models, methods such as backpropagation algorithm, shared weights, and parallel operations were adopted. This article compared different types of deep learning models used in the model, such as CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), and analyzed their performance advantages and disadvantages to help insurance companies choose appropriate model solutions. Model 1 (CNN+vehicle image) has an accuracy of 0.85, a recall rate of 0.80, and an F1 score of 0.82; Model 2 (RNN+text description) has an accuracy of 0.75, a recall rate of 0.72, and an F1 score of 0.73. This article presents an intelligent insurance actuarial model based on machine learning and data mining, which helps to achieve accurate prediction and pricing of insurance actuaries.