Skip to main content

Table 1 Simulation results for Models (7)–(9)

From: Sparse sliced inverse regression for high dimensional data analysis

  

\(\boldsymbol{n}=\textbf{100},\,\boldsymbol{p}=\textbf{150}\)

\(\boldsymbol{n}=\textbf{200},\,\boldsymbol{p}=\textbf{150}\)

Model (7)

Model (8)

Model (9)

Model (7)

Model (8)

Model (9)

Proposed method

TPR

100 (2.3)

99 (0.1)

84 (1.8)

100 (0.0)

100 (0.0)

91 (1.4)

FPR

6.8 (0.5)

11 (0.1)

7.5 (8.0)

0.0 (0.0)

0.0 (0.0)

1.0 (0.1)

Corr

98 (2.3)

94 (0.1)

72 (0.2)

99 (0.0)

99 (1.0)

97 (2.0)

[12]

TPR

96 (1.0)

94 (1.2)

91 (1.1)

98(0.5)

99 (0.5)

99 (2.5)

FPR

6.0 (0.9)

3.6 (0.7)

7.4 (0.1)

3.4 (0.4)

1.1 (0.2)

2.5 (0.3)

Corr

88 (0.9)

86 (1.1)

74 (1.1)

91 (0.5)

92 (0.5)

79 (0.6)

[16]

TPR

95 (0.9)

100 (0.0)

100 (0.6)

100 (0.0)

100 (0.0)

100 (0.0)

FPR

4.9 (0.1)

4.8 (0.1)

3.5 (0.1)

5.9 (0.2)

6.7 (0.3)

4.5 (0.2)

Corr

59 (1.1)

88 (0.5)

79 (0.6)

79 (0.6)

94 (0.2)

87(0.5)

[7]

TPR

98 (0.1)

98 (0.1)

98 (0.1)

99 (0.1)

99 (0.1)

98 (0.1)

FPR

8.3 (1.2)

3.8 (0.8)

23 (1.1)

1.2 (0.4)

0.3 (0.2)

20 (1.1)

Corr

84 (0.9)

89 (0.6)

63 (0.7)

94 (0.4)

96 (0.3)

70 (0.5)

[10]

TPR

89 (1.5)

94 (1.2)

80 (1.2)

98(1.0)

99 (0.7)

96 (0.6)

FPR

0.6 (0.1)

0.6 (0.1)

0.2 (0.1)

0.3 (0.1)

0.3 (0.1)

0.1 (0.1)

Corr

82 (1.4)

85 (1.3)

70 (1.1)

91 (1.1)

93 (1.0)

84 (0.7)

  1. Corr is the correlation coefficient between the true and estimated sufficient predictors; TPR is the true positive rate; FPR is the false positive rate. The mean (standard error), averaged over 200 independent replications, are reported. All entries are multiplied by 100