Fig 1.
Amplitude and time-frequency spectrograms are shown for representative exemplars of the marmoset call types considered in this study. A: Alarm, B: Chirp, C: Loud shrill, D: Phee-2, E: Phee-3, F: Phee-4, G: Seep, H: Trill, I: Tsik, J: Tsik-Ek, K: Twitter.
Table 1.
The number of calls considering each class.
Fig 2.
The effect of training set size on classification performance.
For the sake of visual clarity, the results of OPF using the distance metrics Bray-Curtis and Chi-Square, and SVM using linear and polynomial kernels are excluded.
Table 2.
Classification of all eight classes of vocalizations using different algorithms.
90% of the samples were used for the training set. Time refers to the time required to classify one sample in milliseconds.
Table 3.
Confusion matrix considering the classification of all eight classes of vocalizations using OPF with Manhattan distance and 90% of the samples for training set.
Table 4.
Confusion matrix considering the classification of the principal Phee class into sub-classes using OPF with Euclidean distance metric and 90% of the samples for training set.
Table 5.
Confusion matrix for the classification of the principal Tsik class into sub-classes using OPF with Manhattan distance metric and 90% of the samples for training set.