Application of Bayesian network and regression method in treatment cost prediction

Table 2 Comparison of accuracy of model prediction

Model	Accuracy (%)	MSE	R2
Linear regression model	59.74	13.32	0.58
LASSO regression model	65.78	10.88	0.66
Neural network model	63.45	11.06	0.62
Locally weighted LASSO regression model	85.65	6.38	0.75
Bayesian network fusion local weighted LASSO regression model	89.14	5.36	0.81

Table shows the accuracy, mean square error (MSE) and R-Square (R2) of the prediction results obtained by various regression and prediction methods. MSE is the average of the square of the difference between the predicted value and the true value. The range of R-squared is 0–1. The larger it is, the better the model fitting effect is

ISSN: 1472-6947