About this article
Published Online: May 26, 2023
Page range: 85 - 91
DOI: https://doi.org/10.2478/ijanmc-2022-0019
Keywords
© 2022 Ziwei Ding et al., published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 5.
Figure 6.
Model metrics
| ACC | 96.94% |
| Precision | 96.27% |
| Recall | 96.27% |
| F1 score | 0.9627 |
| AUC | 0.9701 |
Model results
| Forecast | ||
|---|---|---|
| Normal data | Fault data | |
| Normal data | 15611 | 14 |
| Fault data | 14 | 361 |
j.ijanmc-2022-0019.tab.004
| Improved random forest algorithm based on fault ratio |
|
Step1: The training samples were randomly put back from the data set, and were extracted for n times in total to obtain n independent training sets with repeated elements. Step2: The n decision trees are trained on different training sets. Step3: The sample category labels corresponding to N decision trees were analyzed, and the final voting induction was carried out by combining the improved voting decision method based on fault ratio. |
Data variance comparison
| aa_000 | 1.454301e+05 | 1.454301e+05 |
| ac_000 | 7.767625e+08 | 7.724678e+08 |
| ad_000 | 3.504525e+07 | 3.504515e+07 |
| ae_000 | 1.581479e+02 | 1.581420e+02 |
| af_000 | 2.053871e+02 | 2.053753e+02 |
| ... | ... | |
| ee_007 | 1.718666e+06 | 1.718366e+06 |
| ee_008 | 4.472145e+05 | 4.469894e+05 |
| ee_009 | 4.721249e+04 | 4.720424e+04 |
| ef_000 | 4.268570e+00 | 4.268529e+00 |
| eg_000 | 8.628043e+00 | 8.627929e+00 |