Skip to main content

Table 6 SC baseline classification accuracy (%)

From: High level feature extraction for the self-taught learning algorithm

Training

Dictionary size K

Set

100

200

300

500

IS-20

52.5

52.4

51.5

54.1

IS-50

56.3

58.6

58.9

60.9

IS-100

57.0

58.9

60.1

62.3

IS-250

57.4

59.9

61.7

64.0

  1. Bases are learned from the labeled ISMIR training data.