Research Article
Implementation of Process-Based and Data-Driven Models for Early Prediction of Construction Time
Table 6
Model accuracy results for (a) training and (b) validating data by using only data-driven GRNN model (using the 10-fold cross-validation method) [
63].
| (a) Training data | Mean target value for input data | 286.90647 | Mean target value for predicted values | 286.19978 | Variance in input data | 99430.392 | Residual (unexplained) variance after model fit | 2319.0624 | Proportion of variance explained by model (R2) | 0.97668 (97.668%) | Coefficient of variation (CV) | 0.167848 | Normalized mean square error (NMSE) | 0.023323 | Correlation between actual and predicted target values | 0.988279 | Maximum error | 300.05113 | RMSE (root mean squared error) | 48.156644 | MSE (mean squared error) | 2319.0624 | MAE (mean absolute error) | 21.681144 | MAPE (mean absolute percentage error) | 19.471364 |
| (b) Validation data | Mean target value for input data | 286.90647 | Mean target value for predicted values | 285.08662 | Variance in input data | 99430.392 | Residual (unexplained) variance after model fit | 24222.689 | Proportion of variance explained by model (R2) | 0.75639 (75.639%) | Coefficient of variation (CV) | 0.542464 | Normalized mean square error (NMSE) | 0.243615 | Correlation between actual and predicted target values | 0.879069 | Maximum error | 1002 | RMSE (root mean squared error) | 155.6364 | MSE (mean squared error) | 24222.689 | MAE (mean absolute error) | 74.291454 | MAPE (mean absolute percentage error) | 31.79981 |
|
|