Research Article

Application of Boosting Regression Trees to Preliminary Cost Estimation in Building Construction Projects

Table 3

Cost estimation results of each test set.

Number Historical cost
(1,000 KRW)
Neural networksBoosting regression tree
Predicted cost
(1,000 KRW)
Error rate (%)Predicted cost
(1,000 KRW)
Error rate (%)

16,809,4507,704,03413.147,206,7955.84
29,351,71610,015,9067.109,805,6564.85
36,656,2307,251,3178.946,322,1125.02
47,119,4707,128,5130.137,418,3734.20
57,304,7477,978,9909.237,349,1780.61
69,729,3929,516,9462.189,259,1624.83
710,801,8269,817,9999.119,682,11910.37
87,944,3187,246,7638.787,136,77310.17
910,879,00410,136,4316.8310,572,7772.81
107,552,8147,764,3002.807,683,2951.73
118,845,0998,558,5363.248,370,4975.37
1210,690,80010,001,5036.4510,015,2846.32
138,694,7218,258,4525.028,446,7962.85
146,582,6366,810,4063.466,954,5075.65
157,583,6808,312,2169.618,194,2928.05
167,099,2207,955,96612.078,292,38116.81
178,145,1478,604,4445.648,522,0094.63
188,652,8107,853,7659.238,270,1694.42
1910,527,27810,040,0394.639,611,1948.70
206,679,9246,467,3443.187,397,92310.75
218,383,8309,203,8879.788,487,2861.23
227,298,9328,018,2259.858,294,89513.65
237,505,4287,749,0533.257,967,2656.15
247,710,9217,622,0531.157,795,5631.10
256,196,6526,503,0224.945,940,6344.13
268,897,8618,554,4553.868,714,1232.06
277,840,7878,535,6178.868,863,97513.05
288,023,0677,666,8984.446,900,06814.00
297,495,2137,270,8062.997,695,6132.67
307,653,0058,003,2924.587,775,1391.60

MAERs6.055.80