Table 1 Examples of the most common... | Emerald Publishing

Table 1

Examples of the most common prediction tools in civil engineering

Tool	Prediction	Reference	Weakness
Regression analysis	Concrete cost estimation	Tam and Fang (1999)	Inappropriate when describing non-linear relationships, consisting of multiple inputs and multiple outputs (Tam and Fang, 1999)
	Material quantity estimation	García de Soto et al. (2014)
	Building cost estimation	Kim et al. (2004)
Neural networks	Road accidents	García de Soto et al. (2018)	Lose their effectiveness when the patterns are very complicated or noisy, the problem representation and problem structuring are ill defined and training may be trapped in local minima; long computational time (Hegazy et al., 1994); model needs to be retrained when new information is available (Hong et al., 2002)
	Cost estimation of projects	Hegazy and Ayed (1998), Kim et al. (2004)
	Pipe failure prediction	Kerwin et al. (2019)
	Structural health monitoring	Neves et al. (2017)
Case-based reasoning	Building cost estimation	García de Soto and Adey (2015, 2016), Kim et al. (2004)	Limitations to reflect suitable search criteria to index and match depending on previous experience without validating them in a new situation
Case-based reasoning	Building design	García de Soto et al. (2020)
Bayesian networks	Occurrence of road accidents	Deublein et al. (2015, 2013)	Difficulty in developing a Bayesian network with both discrete and continuous variables (Deublein et al., 2013). Current Bayesian network applications handle mostly discrete variables (Hu and Mahadevan, 2018). A vast amount of data may be needed for the learning of the network (Delgado-Hernández et al., 2014)
	Project cost risk analysis	Khodakarami and Abdi (2014)
	Bridge condition modelling	Rafiq et al. (2015)
	Risk assessment	Delgado-Hernández et al. (2014)
	Quantifying schedule risk	Luu et al. (2009)

Tool	Prediction	Reference	Weakness
Regression analysis	Concrete cost estimation	Tam and Fang (1999)	Inappropriate when describing non-linear relationships, consisting of multiple inputs and multiple outputs (Tam and Fang, 1999)
	Material quantity estimation	García de Soto et al. (2014)
	Building cost estimation	Kim et al. (2004)
Neural networks	Road accidents	García de Soto et al. (2018)	Lose their effectiveness when the patterns are very complicated or noisy, the problem representation and problem structuring are ill defined and training may be trapped in local minima; long computational time (Hegazy et al., 1994); model needs to be retrained when new information is available (Hong et al., 2002)
	Cost estimation of projects	Hegazy and Ayed (1998), Kim et al. (2004)
	Pipe failure prediction	Kerwin et al. (2019)
	Structural health monitoring	Neves et al. (2017)
Case-based reasoning	Building cost estimation	García de Soto and Adey (2015, 2016), Kim et al. (2004)	Limitations to reflect suitable search criteria to index and match depending on previous experience without validating them in a new situation
Case-based reasoning	Building design	García de Soto et al. (2020)
Bayesian networks	Occurrence of road accidents	Deublein et al. (2015, 2013)	Difficulty in developing a Bayesian network with both discrete and continuous variables (Deublein et al., 2013). Current Bayesian network applications handle mostly discrete variables (Hu and Mahadevan, 2018). A vast amount of data may be needed for the learning of the network (Delgado-Hernández et al., 2014)
	Project cost risk analysis	Khodakarami and Abdi (2014)
	Bridge condition modelling	Rafiq et al. (2015)
	Risk assessment	Delgado-Hernández et al. (2014)
	Quantifying schedule risk	Luu et al. (2009)