Lasso Regression with Radial Basis Function based Intelligent Music Teaching System

In recent years, information of music education system helps to promote the music teaching by considering music disciplines including nature of music. The music evaluation systems which in exist with Radial Basis Function (RBF) based systems are inefficient solutions for the present difficulties such as lack of teachers and single teaching to realize the teaching resources. In this paper, a generalization technique LASSO regression with Radial Basis Function (Lasso-RBF) is proposed to enhance the performance of intelligent music teaching system. Primarily, the input data is taken from the lakh MIDI dataset and further processed for preprocessing with min-max normalization. Then, hamming window is introduced to extract the features including rhythms, pitches, and melody features. After that, Lasso regression is utilized to select the features to increase the model performance. Finally, the RBF technique is used for predicting the student’s outcomes for teaching music effectively. From the results, proposed LASSO-RBF method attained improved results when compared with existing model such as Gradient Boosted Decision Trees (GBDT) by providing results including accuracy (89.47%), AUG (95.50%), Recall (80.26%), and F1 score (85.92%) correspondingly.