Open AccessSelection of Audio Learning Resources Based on Big Data
Author(s) -
Peng Wang,
Xia Wang,
Xia Liu
Publication year - 2022
Publication title -
international journal of emerging technologies in learning/international journal: emerging technologies in learning
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.454
H-Index - 24
eISSN - 1868-8799
pISSN - 1863-0383
DOI - 10.3991/ijet.v17i06.30013
Subject(s) - computer science , spectrogram , stability (learning theory) , big data , mel frequency cepstrum , collaborative filtering , artificial intelligence , machine learning , multimedia , speech recognition , data mining , recommender system , feature extraction
Currently, audio learning resources account for a large proportion of the total online learning resources. Designing and implementing a method for optimizing and selecting audio learning resources based on big data of education will be of great significance to the recommendation of learning resources. Therefore, this paper studies a method for selecting audio learning resources based on the big data of education, with music learning as an example. First, the audio signals were converted into mel spectrograms, and accordingly, the mel-frequency cepstral coefficient features of audio learning resources were obtained. Then, on the basis of the conventional content-based audio recommendation algorithm, the established interest degree vector of target students with respect to music learning was expanded, and a collaborative filtering hybrid algorithm for audio learning resources that incorporates the interest degrees of neighbouring students was proposed, which effectively improved the accuracy and stability in the prediction of students’ interest in music learning. Finally, the experimental results verified the feasibility and prediction accuracy of the proposed algorithm.