Context Adaptation of Fuzzy Inference System-Based Construction Labor Productivity Models

Table 1

Base context-specific CLP models: features, structure, and model parameters.

Features, FIS structure, and model parameters	Context
	1	2	3	4
	Concreting, industrial buildings	Concreting, warehouse buildings	Concreting, high-rise buildings	Concreting, institutional buildings
Number of input features	16	7	8	11
Number of data instances	23	16	28	25
Model features	,	,	,	,
Fuzzification coefficient	1.5	2.5	2.0	2.0
Number of rules	6	7	6	7
Input aggregation operator	PROD	MIN	PROD	PROD
Implication method	PROD	MIN	PROD	PROD
Rule aggregation operator	MAX	SUM	PROBOR	PROBOR
Defuzzification method	MOM	BISECTOR	CENTROID	BISECTOR
Accuracy (RMSE)	1.162	0.467	0.992	0.671

Note. Model features represent: -crew size, -craftsperson on-job training, -crew composition, -cooperation among craftspersons, -team spirit of crew, -level of interruption and disruption, -fairness of work assignment, -location of work scope (distance), -location of work scope (elevation), -congestion of work area, -fairness in performance review of crew by foreman, -site congestion, -treatment of foremen by superintendent and project manager, -uniformity of work rules by superintendent, -out-of-sequence inspection, -safety training, -project safety administration and reporting, -oil price fluctuation, -natural gas price, -concrete placement technique, -structural element type, -direct work proportion, -preparatory work proportion, -tools and equipment proportion, -material handling proportion, -travel proportion, and -personal proportion, -CLP.