Model performance assessment in resilient modulus modelling: a multimodel approach

Ikeagwuani, Chijioke Christopher; Nwonu, Donald Chimobi

doi:10.1080/14680629.2020.1753100

Cited by 14 publications

(4 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…3 for PC 1 and PC 2, which suggests that there is a pattern in the data. This preclusion of a random distribution is an indication of heterogeneity of variances in the data, which is a common observation for the LTPP data utilized in this study [ 25 ]. This lends credence to the adoption of a multivariate method for the descriptive analysis, which is independent of any prior data conformance in the original data.…”

Section: Resultsmentioning

confidence: 99%

“…Furthermore, the class of soils from which the data was obtained includes the American Association of State Highway and Transportation Officials (AASHTO) soil classes A1 to A3, which represent the coarse-grained soils used as unbound granular materials for road pavements. The spatial variability in the extracted soil properties was incorporated by considering different strategic highway research program numbers as reported in a previous study [ 25 ]. The spatial variability was ensured within each of the different US counties that include Wisconsin, Iowa, Delaware, Kansas, Oklahoma, Michigan, Louisiana, Utah, North Dakota and Washington, from which the data were acquired.…”

Section: Methodsmentioning

confidence: 99%

“…The compaction characteristics include the moisture content (Φ) and dry density (γ), while the stress state are the nominal maximum axial stress and confining pressure ( ). The data were manually sorted and pruned before usage to minimize the occurrence of data errors often associated with the LTPP data [ 25 ]. The basic descriptive statistics of the soil data is given in Table 1 .…”

Section: Methodsmentioning

confidence: 99%

“…The most recent studies on prediction based on various AI methods strongly highlight the need for hybrid and more thorough AI approaches to improve the prediction accuracy in comparison with routine AI methods, like the ANN. This is pertinent because the spatial variability of soil properties has been shown to have a great influence on the predictions of the of coarse-grained soils [ 25 ] . Heidarabadizadeh et al [ 26 ] applied two AI methods, which include the hybrid support vector machine and colliding bodies optimization for the prediction of the of unbound subbase materials.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Statistical analysis and prediction of spatial resilient modulus of coarse-grained soils for pavement subbase and base layers using MLR, ANN and Ensemble techniques

Ikeagwuani

Nwonu

2022

Innov. Infrastruct. Solut.

Self Cite

View full text Add to dashboard Cite

Prediction methods for the estimation of the resilient modulus( ) of unbound granular materials continues to be fraught with challenges associated with the in situ dependency of the existing methods, which makes them unreliable when considering the spatial variability of soil properties. Moreover, the artificial intelligence methods, which prove to be superior for prediction, indicate that more recent and hybrid machine learning methods are more apt. Consequently, a comparative analysis was executed in the present study using advanced statistical approaches and a simple routine method for the prediction of the of coarse-grained soils used as unbound granular materials in highway pavements. This study considered also the spatial variability of the routine soil properties based on the long-term pavement performance database. The result of the descriptive analysis conducted using principal component analysis revealed that the variation in is due to three principal factors, which include the effect of moisture on the fines content, effect of coarse particles on the compaction characteristics and the effect of the soil stress state. Subsequently, prediction of the using multiple linear regression was unreliable based on residual analysis and multi-collinearity problems. Thus, predictive modeling of was then executed using artificial neural network (ANN) and ensemble techniques (random forest and gradient boosting machine), which resulted in very high R 2 values above 0.9, considered apt for MEPD. The ANN model was found to provide a superior prediction performance than the ensemble techniques.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%