Network Audio Data and Music Composition Teaching Based on Heterogeneous Cellular Network

Table 3

Comparison of the test results of the original classification system, decision tree model, and random forest model using the TUT data set.

Features and decision models	Flat measurement index
	Ambient sound accuracy (%)	Environmental sound recall rate (%)	Music accuracy (%)	Music recall rate (%)
7-dimensional features + 3sigma judgment	76.19	38.01	72.22	93.14
1572-dimensional features + decision tree decision	84.60	96.07	97.54	89.90
46-dimensional features and random forest model after feature selection	94.61	94.06	96.58	96.90