Table 4 The estimated number of problems for five real datasets from published usability studies
n* | j** | naïve | Good-Turing | double deflation | LNBzt | matrix-based | |
|---|---|---|---|---|---|---|---|
EDU3D | 20 | 119 | |||||
\( \hat{m} \) | 120 | 121 | 122 | 155 | 152 | ||
95%CI | 117–121 | 118–125 | 120–129 | 132–195 | 135–167 | ||
MACERR | 15 | 145 | |||||
\( \hat{m} \) | 156 | 178 | 184 | 449 | 382 | ||
95%CI | 146–160 | 171–207 | 192–245 | 256–1301 | 346–440 | ||
MANTEL | 76 | 30 | |||||
\( \hat{m} \) | 30 | 30 | 30 | 31 | 30 | ||
95%CI | 30–30 | 30–30 | 30–30 | 31–35 | 30–37 | ||
SAVINGS | 34 | 44 | |||||
\( \hat{m} \) | 44 | 44 | 44 | 46 | 45 | ||
95%CI | 44–45 | 44–45 | 44–45 | 42–50 | 44–51 | ||
INFPUMP | 34 | 107 | |||||
\( \hat{m} \) | 107 | 107 | 107 | 122 | 120 | ||
95%CI | 107–108 | 106–108 | 106–108 | 110–136 | 112–143 |