Skip to main content
. 2020 Nov 9;3(4):77. doi: 10.3390/mps3040077

Table 1.

Results of five regression models predicting FEV1 using the best features.

Regression Model R2 RMSE
Linear 74.98 ± 7.45 0.3781 ± 0.06380
Lasso
(α = 0.0001)
74.99 ± 7.45 0.3801 ± 0.0519
Ridge
(α = 0.4)
75.03 ± 7.37 0.3780 ± 0.0639
Elastic Net
(α = 0.001)
75.00 ± 7.41 0.3781 ± 0.0640
Bayesian Ridge 75.01 ± 7.42 0.3780 ± 0.0639

The models were developed using the four best features (height, sex, and weight at age 18 and FEV1 at age 10) as predictors of FEV1. Here, R2 = average goodness-of-fit measure for regression models represented as a percentage and RMSE = average root mean squared error.