Adaptive Data Boosting Technique for Robust Personalized Speech Emotion in Emotionally-Imbalanced Small-Sample Environments

. 2018 Nov 2;18(11):3744. doi: 10.3390/s18113744

Algorithm 1 Insufficient Data Reinforcement

Input:

T D S (1 \dots N)

—Target User Dataset

D S (1 \dots M)

—Initial Multiple User Dataset

FN—Number of Features

C—Number of Classes

Output: S (1...K)—Selected Similar Emotional Speeches Dataset

for i = 1 to N do

T F e a t u r e V e c t o r_{i}

= extractFeatures(

T D S_{i})

;

{T E m o L a b e l}_{i}

= getLabel(

T D S_{i})

;

end

for i = 1 to C do

cnt = 0;

for j = 1 to FN do

for k = 0 to N do

T E m o L a b e l_{k}

= i then

T s u m_{i j} = T s u m_{i j} + T F e a t u r e V e c t o r_{k j}

;

cnt ++;

end

T m e a n s_{i j}

T s u m_{i j}

/cnt;

end

for i = 1 to C do

for j = 1 to M do

I F e a t u r e V e c t o r

= extractFeatures(

I D S_{j})

;

D i s t a n c e s_{i j}

= d(

T m e a n s_{i}, I F e a t u r e V e c t o r);

End

end

S = Sorting (

D i s t a n c e s, I D S)

;

Return S;