From: Evaluation of the effectiveness and efficiency of state-of-the-art features and models for automatic speech recognition error detection
Feature set
Accuracy (%)
F-measure (%)
Correct
Substitution
Insertion
Average
MLP
83.97
90.84
57.84
01.70
82.41
ULSTM
83.39
90.59
49.90
01.69
80.76
BLSTM
83.41
52.13
03.14
81.21
V-RNN
85.10
91.58
57.81
04.10
83.05