A Gated Recurrent Unit (GRU) Model for Predicting the Popularity of Local Musicians

Purpose: Popular musicians are among the most admired people in the world, and music is one of the features of modern culture that is most universally appreciated. Why one musician is well-liked while others are not is frequently a very tough question to answer. Survey shows most local yet talented musicians were not explored but lost due to unwillingness of music promoters to promote them. This study aims at using Gated Recurrent Unit (GRU) model in predicting the rise and popularity of local musicians by considering some established metrics for the evaluation of local musicians, alongside some characteristics/features. Design/Methodology/Approach: Using the Kaggle dataset of local musicians’ performance characteristics and historical data, the study analyzes certain features of local musicians and predicts their possible future rise, using a GRU Deep Learning Model. Findings: The result shows a high degree (70.1%) of accuracy with a Mean Square Error (MSE) of 0.0069, between the predicted and actual performance. The work shows that the developed model can predict the possible future rise and popularity of local musicians. Research Limitations: Only a few available original data (relating to local unpopular musicians) were extracted and due to time constraints in implementing the work, the authors could not get down to the community to extract more from local musicians around. Practical Implication: The model can be used by music promoters to decide on sealing contract agreements with local/upcoming musicians both for their sustainability and promotability. Originality/Value: While many predictive works were done focusing on music popularity, the song hits rate, and so on, the few that centered their aim on musicians’ popularity only analyzed established musicians. This work however ventures into the analysis of the possible elevation of local musicians by considering certain parameters that relate to their songs.