Research Article

Optimising the Selection of Input Variables to Increase the Predicting Accuracy of Shear Strength for Deep Beams

Table 8

A comparison between the suggested models’ efficiency in the presence of GAITH.

ModelDistribution of data in terms of relative error for each model
Training phaseTesting phase
5%10%15%20%5%10%15%20%

81.6795.0096.6799.4432.6157.6172.8383.70
63.3377.7889.4493.8931.5246.7463.0471.74
26.6745.5661.1170.5619.5738.0450.0061.96
33.8960.5677.2286.6720.6544.5760.8770.65
82.2295.5697.2299.4433.7062.7877.9186.78
60.5682.7891.6793.8927.1753.2665.2275.00
25.5645.0061.6770.5619.5738.0452.1761.96
37.2265.0082.2292.2221.7443.4860.8771.74
80.5693.8997.7899.4430.4360.8777.1785.87
61.1183.8990.0092.2228.2650.0061.9673.91
27.2245.5660.0071.6721.7439.1348.9163.04
35.5665.5684.4491.6721.7436.9657.6171.74
66.6791.1197.2298.8933.7060.8778.2685.87
55.5678.8989.4492.7829.3551.0967.3975.00
25.0041.1161.1172.7818.4840.2251.0963.04
51.6781.6792.7897.2222.8348.9166.3072.83
23.3342.2261.6775.0016.3039.1352.1768.48
57.2274.4485.0090.5618.4835.8754.3564.13
16.1133.3343.3356.1114.1329.3542.3957.61
36.6765.5683.3392.2222.8342.3959.7871.74