Hyper Parametric Analysis of Multi-Layer Perceptron for Music Genre Classification
Nipun Sharma, Swati Sharma, T. Prabhu, E. Suganya, M Asha, Sandhya Dass · 2024
Machine Learning has played a pivotal role in enhancing user experience in the field of entertainment. A gamut of music related software and applications incorporate machine learning at its core to provide an exceptionally unique user experience. An important aspect of design of such a recommender system is the classification of music. Deep Learning has proven to be really efficient in classification problems of music. However extensive parametric analysis of deep learning methods like Multi-Layer Perceptron (MLP) can yield accuracy of up to 84.1 % and it can give researchers a greater idea of the tuning they can use for quick and efficient classification. This paper presents an extensive analysis of hyper parameter tuning of MLP which can be used as a benchmark for designing more accurate system design.