Table 8.
Comparison with related work in IoT for Healthcare
| Features | Learning Models | Training dataset | Prediction Accuracy (Testing) | ECG signal processing |
|---|---|---|---|---|
| [23] | Modified deep convolutional neural network | UCI heart disease dataset (303 datapoint) | Two level :93% | Not done |
| Three level:- | ||||
| [25] | MSSO-ANFIS | UCI heart disease dataset | Two level :98.79% | No ECG sensor mentioned. |
| Three level:- | ||||
| [9] | decision tree classification algorithm based on Iterative Dichotomiser 3 (j48 classifier) | UCI heart disease dataset | Two level :91.48% | No ECG sensor mentioned |
| Three level:- | ||||
| [10] | Random Forest classifier | UCI heart disease dataset (270 datapoint) | Two level :99% | implements an edge computing technique to find out three slope characteristics of the ST wave. |
| Three level:- | ||||
| [24] | Random Forest | consist of 191 records of the patient | Two level : 93% | No ECG sensor mentioned |
| Three level:- | ||||
| [30] | SVM | UCI heart disease dataset | Two level :97% | No ECG sensor mentioned |
| Three level:- | ||||
| Proposed | Stacking Classifier | UCI heart disease dataset (920 datapoint) | Two level :91% | QRS complex detection using pan-tompkins algorithm and slope measurement of the ST segment. |
| Three level:80.4% |