Results of the VSL models on test sets
| Number of firms | Input sequence length | Accuracy | 95% C.I. | ROC-AUC |
|---|---|---|---|---|
| 110k | all lengths | 0.835 | [0.829, 0.832] | 0.912 |
| 55k | 6-7-8-9 | 0.850 | [0.828, 0.842] | 0.927 |
| 43k | 3-4-5 | 0.834 | [0.816, 0.828] | 0.908 |
| 54k | 3-5-7-9 | 0.836 | [0.812, 0.827] | 0.910 |
| Number of firms | Input sequence length | Accuracy | 95% C.I. | ROC-AUC |
|---|---|---|---|---|
| 110k | all lengths | 0.835 | [0.829, 0.832] | 0.912 |
| 55k | 6-7-8-9 | 0.850 | [0.828, 0.842] | 0.927 |
| 43k | 3-4-5 | 0.834 | [0.816, 0.828] | 0.908 |
| 54k | 3-5-7-9 | 0.836 | [0.812, 0.827] | 0.910 |
Note(s): Each model has been trained and tested by using the number of firms in the first column. The table reports the accuracy of the best model (over cross-validation) on the bankruptcy binary classification, followed by the confidence interval of accuracy over the results of cross validation and the ROC-AUC score of the best performing model
Source(s): Table created by the authors
Sharing content requires targeting cookies to be enabled. Please update your cookie preferences to use this feature.