audio segmentaion with bic

i should write the code for an article there is a guidance in the article but i can't understand it. can somebody please help me with it(i'm good in writing codes about video and image but i've never worked on audio before)

here is what article said:

Segmentation of the audio stream was performed using the
Bayesian information criterion (BIC). Bottom-up agglomerative
clustering is based on a variant of the BIC criterion to select the
two closest clusters, where clusters are modeled using 16 component
Gaussian mixture models [17]. Each audio cluster is characterized
by a unique 528-dimensional feature vector gathering scaled
versions of the 33-dimension mean vectors of the 16 Gaussian components.

[17]“Speaker diarization using bottom-up clustering based on a parameterderived distance between adapted gmms,”