Table 6 Performance comparisons of our work with traditional machine learning approaches, in terms of RMSE
From: Stacked LSTM based deep recurrent neural network with kalman smoothing for blood glucose prediction
Name of the model/approach | Mean RMSE for six patient |
|---|---|
Linear regression | 19.62 |
Elastic net regression | 20.15 |
Gradient boosting Tree | 20.73 |
Huber regression | 19.93 |
Lasso regression | 20.22 |
Random forest | 21.05 |
Ridge regression | 19.62 |
SVR with linear kernel | 19.72 |
SVR with radial basis kernel | 19.53 |
Proposed model (without KS) | 18.57 |
Proposed model (with KS) | 5.89 |