Research Article

Weighted Domain Transfer Extreme Learning Machine and Its Online Version for Gas Sensor Drift Compensation in E-Nose Systems

Table 3

Comparisons of DAELM and WDTELM using different numbers of selected samples.

Batch number DAELM (%)WDTELM (%)

66.778.185.585.193.292.882.882.584.090.295.795.7
76.290.3 94.0 96.3 96.3 96.9 75.8 83.0 90.5 90.1 92.4 92.4
100 100 100 100 100 100 87.0 92.5 100 100 100 100
99.4 99.4 99.5 100 100 100 94.4 99.5 99.5 98.9 99.5 98.9
54.6 59.3 73.4 75.7 77.5 80.9 85.5 88.7 94.1 95.3 96.5 98.2
62.6 60.4 64.3 68.3 72.7 74.1 57.5 64.9 79.2 78.1 80.6 87.0
69.4 94.8 95.5 95.9 96.5 96.9 71.4 77.2 88.1 94.6 95.9 97.3
100 100 100 100 100 100 92.1 93.2 95.5 99.1 98.9 99.8
37.0 38.8 41.5 47.9 51.6 52.1 55.8 52.0 58.6 62.6 57.7 62.2

Ave.73.980.183.885.587.688.276.981.587.789.990.892.4