Skip to main content

Table 2 Testing results for different basic variants of the statistics.

From: The comparative analysis of statistics, based on the likelihood ratio criterion, in the automated annotation problem

Statistic

N 1

N 2

N all

P(1) + P(2)

P(1) + P(+)

η[1,1]

118

102

220

0.113

0.188

η[1,0]

130

104

234

0.123

0.201

η[0,1]

136

106

242

0.129

0.208

η[0,0]

164

103

267

0.152

0.232

T(1)[1,0]

213

83

296

0.191

0.260

T(2)[1,1]

172

124

296

0.162

0.256

T(2)[0,1]

216

103

319

0.196

0.281

T(2)[1,0]

180

141

321

0.171

0.277

T(1)[0,0]

200

124

324

0.185

0.283

T ^ ( 2 ) MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xH8viVGI8Gi=hEeeu0xXdbba9frFj0xb9qqpG0dXdb9aspeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGacaGaaiaabeqaaeqabiWaaaGcbaGafmivaqLbaKaadaahaaWcbeqaaiabcIcaOiabikdaYiabcMcaPaaaaaa@2FE5@ [0,1]

189

136

325

0.178

0.282

T ^ ( 2 ) MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xH8viVGI8Gi=hEeeu0xXdbba9frFj0xb9qqpG0dXdb9aspeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGacaGaaiaabeqaaeqabiWaaaGcbaGafmivaqLbaKaadaahaaWcbeqaaiabcIcaOiabikdaYiabcMcaPaaaaaa@2FE5@ [0,0]

234

111

345

0.213

0.304

T ^ ( 2 ) MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xH8viVGI8Gi=hEeeu0xXdbba9frFj0xb9qqpG0dXdb9aspeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGacaGaaiaabeqaaeqabiWaaaGcbaGafmivaqLbaKaadaahaaWcbeqaaiabcIcaOiabikdaYiabcMcaPaaaaaa@2FE5@ [0,0]

205

148

353

0.193

0.307

T ^ ( 2 ) MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xH8viVGI8Gi=hEeeu0xXdbba9frFj0xb9qqpG0dXdb9aspeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGacaGaaiaabeqaaeqabiWaaaGcbaGafmivaqLbaKaadaahaaWcbeqaaiabcIcaOiabikdaYiabcMcaPaaaaaa@2FE5@ [1,1]

142

217

359

0.148

0.294

T ^ ( 2 ) MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xH8viVGI8Gi=hEeeu0xXdbba9frFj0xb9qqpG0dXdb9aspeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGacaGaaiaabeqaaeqabiWaaaGcbaGafmivaqLbaKaadaahaaWcbeqaaiabcIcaOiabikdaYiabcMcaPaaaaaa@2FE5@ [1,0]

164

224

388

0.167

0.321

q

393

158

551

0.354

0.502

S And

332

174

506

0.304

0.489

  1. The table shows the testing results of the different statistics for n all = 9236 'predictions'. In n q = 1176 cases the "predicted word" belonged to the sequence, in n all - n q = 8060 it did not. N1 denotes the number of false negatives, N2 is the number of false positives. Nall= N1 + N2, P(1) = N1/n q , P(2) = N2/(n all - n q ) and P(+) = N2/N2 + (n q - N1)). See the main text for the description of the variants.