“…It is shown that the performance of the 32-sensor and 10-sensor schemes were comparable (67% vs. 69%), however, the six-sensor scheme performed significantly better than its counterparts with 85% classification accuracy. Combined data 31, 5,23,6,28,26,29,9,20,18 It is interesting to note that although the six-sensor scheme trailed the other two schemes in the model cross validation, it outperformed the other two schemes considerably in the model validation. The potential reason for this performance discrepancy between the six-sensor scheme and 32-or 10-sensor schemes may lie in the fact that the four sensors (sensor 5, 6, 23, and 31) that are more sensitive to the water vapor were excluded from these six sensors.…”