An artificial neural network model to relate organisation characteristics and delivery methods of construction projects

Fig. 1:

Fig. 2:

Fig. 3:

Fig. 4:

Fig. 5:

Fig. 6:

Cronbach’s alpha coefficient for reliability analysis

Construct	Cronbach’s alpha
Managerial expertise	0.82
Financial stability	0.84
Technical capability	0.80
Project experience	0.83
Contextual factors	0.85
Overall questionnaire	0.85

Statistical summary of the collected data

Variable	Mean	Standard deviation	Minimum	Maximum
Managerial expertise	3.8	0.7	2	5
Financial stability	3.5	0.8	1	5
Technical capability	4.0	0.6	2	5
Project experience	3.7	0.7	2	5
Contextual factors	3.6	0.8	1	5

Comparing the accuracy of data mining models

Models	Accuracy criterion
Naive Bayes	48.11
Decision tree	26.42
Random forest	13.21
SVM	13.21
Selected neural network model	76.42

Performance comparison of different models

Model	R2	MAPE (%)
ANN	0.82	12.5
Linear regression	0.65	18.3
Decision tree	0.70	15.2

Example of normalised data

Respondent ID	Managerial expertise	Financial stability	Technical capability	Project experience	Contextual factors
1	0.75	0.80	0.60	0.70	0.65
2	0.60	0.55	0.75	0.65	0.70
…	…	…	…	…	…
354	0.80	0.70	0.85	0.75	0.80

Connection coefficients of nodes in the output layer (best model)

Node	Output Node 1	Output Node 2	Output Node 3	Output Node 4	Output Node 5	Output Node 6
Node 1	–1.454	–1.818	–1.712	–2.103	–2.073	–4.525
Node 2	–3.075	–3.961	–2.565	–11.451	–10.190	2.895
Node 3	2.199	–4.315	–0.647	0.256	2.389	–1.400
Node 4	–2.566	–2.195	0.493	6.947	16.573	9.320
Node 5	–3.351	–3.045	–1.206	–2.245	–5.482	–3.947
Node 6	9.765	10.638	2.272	–3.172	–14.231	–23.596
Threshold	–4.720	–4.891	–2.454	–3.010	–5.765	–4.569

Overview of input variables

Variable	Type	Description	Indicators
Managerial expertise	Numerical	Measures the level of expertise and experience of the managerial team	Project management plan, supervision team management, coordination of planning, managerial support, adoption of new methods, and continuous monitoring and control
Financial stability	Numerical	Assesses the financial health and stability of the organisation	Timely payment, cost planning, planning and control system, project pricing, economic justification, and financial resource creation
Technical capability	Numerical	Evaluates the technical skills and capabilities of the organisation	Technical foundation, reporting systems, and optimisation programmes
Project experience	Numerical	Reflects the organisation’s experience with similar projects	Project management experience, training and development, and human resources satisfaction
Contextual factors	Numerical	Considers external factors that might impact project delivery	Infrastructure, legal obstacles, training, and culture development

Connection coefficients of nodes in the hidden layer (best model)

Variable	Node 1	Node 2	Node 3	Node 4	Node 5	Node 6
Managerial	4.814	5.094	4.814	–4.811	–4.052	–3.215
Financial	9.256	8.379	9.256	–9.498	–8.366	–5.841
Background	5.696	3.822	5.696	–4.698	–3.133	–2.998
Optimisation	2.981	2.587	2.981	–2.511	–2.684	–1.871
Energy efficiency	0.581	0.510	0.581	–0.940	0.019	–0.506
Bias	–7.437	2.136	–7.437	1.524	9.801	–3.991

PDMs

Escrow (one factor)

Percentage

Unit price

Price list (without aggregates or with aggregates)

MC

BOT

Purchase and installation

Design, purchase (procurement), and execution (Engineering, Procurement, and Construction – EPC) or design and construction

Project delivery systems

Multifactorial method (multiple prime)

Construction manager method

DB method