Skip to main content

Table 1 Results of Pearson’s correlations (R) and root mean squared errors (RMSE) comparing predictions under the machine learning approaches proposed and the experimentally measured RWC values

From: Instantaneous and non-destructive relative water content estimation from deep learning applied to resonant ultrasonic spectra of plant leaves

Number of ultrasonic signalsMethodRRMSE
1960Random forest0.84000.0591
1960Convolutional neural network0.92250.0407
280Random forest0.84530.0585
280Convolutional neural network0.92280.0406