Research on piano music recommendation algorithm based on recurrent time convolutional network

Rui Zhao
College of Music Education, Xi'an Conservatory of Music, Xi’an, China, 710061
Mengqian Lin
Shanghai Modern Academy of Family Education.,Shang’hai,China, 200063

Abstract:

This paper studies piano music recommendation algorithm based on recurrent time convolutional network (RTCNN) and it tries to identify user's music preferences by the style of music and genre. The RTCNN is a novel deep learning framework which has a series of layers and units that work in sequence. They are designed in such way that they can learn complex sequential patterns from raw data with minimal supervision from experts or domain experts. After finishing the experiments, six most popular genres in China such as pop, jazz, classical, R&B/hip-hop, rock and Chinese music were identified through the shortlisting process with an accuracy rate of 73.88%.The experimental results of this paper show that the language, social, cognitive and behavioral problems of autistic children under music therapy under intelligent health monitoring have been significantly improved, and their emotional response ability has improved the most, increasing by 34%. Communication ability was next, increased by 20.3%, motor coordination ability increased by 20%, and cognitive ability improved by 11%. It can be explained that the four aspects of language, social interaction, cognition and behavior have been significantly improved after music therapy.