Melody Classification based on Performance Event Vector and BRNN

TL;DR

This paper presents a melody classification model using Performance Event Vectors and Bidirectional RNNs, demonstrating effective results on multiple datasets and analyzing hyper-parameter impacts.

Contribution

The paper introduces a novel approach combining Performance Event Vectors with BRNNs for melody classification, advancing the methodology in music information retrieval.

Findings

Achieved high accuracy on development and Wikifonia datasets

Analyzed hyper-parameter effects on model performance

Generated multiple predictions for evaluation robustness

Abstract

We proposed a model for the Conference of Music and Technology (CSMT2020) data challenge of melody classification. Our model used the Performance Event Vector as the input sequence to build a Bidirectional RNN network for classfication. The model achieved a satisfying performance on the development dataset and Wikifonia dataset. We also discussed the effect of several hyper-parameters, and created multiple prediction outputs for the evaluation dataset.