Calibration and verification: Laboratory calibration of the DFM
temperature sensors was conducted with the data from the first
set of the temperature sensors. The thermometer observations
were regressed against the DFM temperature sensor readings.
The resulting linear equation was termed the ‘‘laboratory calibration
line (LCL)”. The LCL was used to modify the temperature readings
from the second set of individual DFM temperature sensors.
Goodness of fit for the modification of temperature readings
against the thermometer readings was carried out with coefficient
of determination (r
2
), the D-index of agreement (Eq. (4)), while
accuracy of modification was done using mean absolute error
(MAE; Eq. (5)) and root means squared error (RSME; Eq. (6))
(Walther and Moore, 2005). According to Willmont (1982), any
model with r
2 that approaches 1 has a good performance, and a
D-index close to 1 as well as MAE close to 0 illustrates good
agreement between measured and modified values.