A LightGBM-Based Pricing Method for Healthcare Data

In the process of deepening development of digital economy, healthcare data have become key resources due to their multi-dimensional information and application value, but their pricing faces challenges such as privacy protection, data heterogeneity and scenario diversification. Traditional pricing models are difficult to effectively capture high-dimensional nonlinear relationships, and there is an urgent need for breakthroughs in intelligent algorithms. This paper proposes a healthcare data pricing model based on LightGBM. In the empirical part, 397 transaction records of healthcare data are used, and RF, XGBoost and other models are selected as benchmark models for comparative experiments. The effectiveness of the models are evaluated through root mean square error (RMSE) and mean absolute error (MAE). The results show that LightGBM reduces RMSE and MAE by 5.40% and 11.68% respectively compared to the second-best model, significantly better than other algorithms. Research has shown that the LightGBM model can efficiently handle the sparsity and nonlinear features of data, providing a high-precision and low-error solution for the pricing of healthcare data.