Language Recognition Using Latent Dynamic Conditional Random Field Model with Phonological Features

Table 8

Language classifier error using different speech feature sets (averaged across 15 pairs).

Feature	Error (%)
PF	15.70
SDPF	19.26
MFCC	23.87
PLP	17.39
SDC	21.94
SDC + PLP	19.04

The least error result is displayed in bold.