.png){kind=icon}
.png){kind=icon}
Return
AI-Driven predictive modelling for health insurance pricing with secure cloud deployment
.png){kind=icon}
PRE
AI
R
B
V
C
N
G
DOI：10.1080/17509653.2026.2668574.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Various factors, such as demographic characteristics, medical history, and claims history influence health insurance pricing. Predicting pricing strategies and maintaining data security are important aspects of health insurance. This study proposes the factors influencing health insurance data and develops predictive models using Artificial Intelligence (AI) techniques to understand pricing strategies in the health insurance sector. This predicted AI model is encrypted and stored in a cloud database to ensure data security and management. Initially, the US Health Insurance Dataset is utilized, Hot Deck Imputation is applied for missing values, and Cook’s distance for outlier detection to increase the reliability of the data. The discrete cosine transform key features are extracted for prediction using the transform. Attention in an Echo State Network (ESN) increases the accuracy of pricing. Brakerski-Fan-Vercauteren encryption ensures sensitive insurance data is protected, and private cloud deployment is made possible by encryption, which secures the AI model. The experimental results proved the efficiency of the method for price predictions with a 0.07235 Mean Absolute Error and R2 score of 99.57%, while imposing security with an encryption time of 26.65 seconds for 1000MB of model data, thus providing a robust and scalable solution for health insurance pricing.
Keywords:
Health insurance pricing
machine learning
echo state networks
attention mechanism
homomorphic encryption
cloud security
G22
C45
M15
Journal
IF:
2.6
Papers:
237
Citations:
739
Organization
Cited Papers