Table 3 DP-SMSN algorithm in 3DRoad data set.

From: Differentially private density estimation with skew-normal mixtures model

k

2

3

4

5

6

7

8

DPGMM

AIC

4,350,524

4,242,447

4,103,088

4,057,006

4,006,059

3,957,658

3,882,209

BIC

4,350,733

4,242,766

4,103,516

4,057,544

4,006,707

3,958,415

3,883,077

DP-MSNM

AIC

4,194,876

4,148,930

4,139,086

3,995,200

3,941,379

3,898,330

3,904,047

BIC

4,195,150

4,149,348

4,139,646

3,995,903

3,942,225

3,899,318

3,905,178

  1. For each component \(k(k = 2, \ldots , 8)\), the Akaike (AIC) and Bayesian (BIC) information criteria, appear in two models.