Research Article
Prediction of Bridge Component Ratings Using Ordinal Logistic Regression Model
Table 8
Performance improvements by ordinal model compared with multiple linear model.
| Model Statistics | 2012 | 2013 | 2014 | Deck | Superstruc. | Substruc. | Deck | Superstruc. | Substruc. | Deck | Superstruc. | Substruc. |
| Sum of Absolute Residuals | 13.5% | 21.6% | 7.56% | 13.3% | 19.2% | 4.71% | 12.9% | 19.6% | 5.25% |
| Sum of Residual Squares | 21.0% | 27.0% | 12.4% | 21.4% | 26.2% | 9.7% | 22.0% | 26.8% | 11.0 % |
| Bias | 5.61% | 6.15% | 4.43% | 6.66% | 7.01% | 4.36% | 5.7% | 5.28% | 4.71% |
| Accuracy | 1.25% | 2.79% | 0.51% | 2.4% | 3.4% | 0.17% | 1.26% | 1.64% | 0.43% |
| Exact Forecasts | 9.33% | 19.4% | 5.93% | 8.01% | 15.6% | 1.59% | 6.46% | 16.0% | 1.5% |
|
|